Slavery and social death: a comparative study 9780674810822, 9780674810839, 9780674744141

Citation preview

le

CO Co oo a ON , LO a lhl Oe en. CO eae

OCS Oe OOOOGs a ,hlm,Utdws es Lo aS a ee Se a a) Se ooFe OO

ee ee eee eee ee ee ee eee Ue PES ONAL MnO ARIS RUI CLOUD NRTA EMER Tan Mn CULO TMM UT UNIT CHC aca nanoRCRI MN UHe KMUn On CoM CUCM UTR a Gen Cn CcoC Cum CU MM nn ONC ete tn Tier

OOO MOC A ONG UU UB AAS aE MOO OUTS OA OMA a UVES ORIG AGE a.

ITS ee AGS AUTOS STIS NSE UCDO SUNRISE GES CRO ACES UN NEUEN AS AORN ERE MCR SUS ECU a ONE MAND SUC UUM UU RON UU Ge AADHIHalai

ITUS COS HENS ROT SECOCO SIE SD ZS SRC GR BIIR UEESSN CSOD ISON EOD UE UU NSE RES UGS EGCU UTIENE SINACONTO SSO ROT ESP MEN SECA GT GHA RUS UU CCA UTROUGE SC TS OURS OE USI GSIRCE SSSI OIG NS UR AARG SAUNA AUS SEI AI OS ED Sl TOI RO Sc SOTASE AD CONUS USN SI UECTE UAMENUS USSR ENS ISIS RC SA SU GS IOIUUG US AOR ASG AURIS, SOS SUE I NEGUS SSSI TOLER COON ORES CCI SIRS ONE OSE ORE SOON: OE ROU BO GUS GETTIN SSO NRSC AE. SRR O MEN CRN AOS SIIZH USE RAT a TSG SIEGE SU OU AOS AE CR UO ESSN SEATS CORALS NEU EC ER HSS UNG CAA SS a UTR VTS RN ACO ESSER UT

eee ee ee eT ee De oo Oe a ve a -— SO SOOT ak oe a EIKO Gs) Glee OOOe es I Ien Ca a.. GGT A Cna OO BSS AECL RUINS SULA AZ NE. Ge SER SIS SS US RG OC CO CRE OE et

TSE TS SGD SCOOTER UD CASA SSLRU OURS SCHIST ANA NUR RAID SSG UU ONSEN UGC OU AOA RUA SANIT TOSSRO UUOa ARAN UG HIE OND CiaOA DeeZEN RU REAR Waa STH ARETE ERVINCU aes CRATE NCTE DE rE CONG EN CHOU RURCNA TAR BE USEEIOHMS HCPA MS GON RRS Cera AI

ASSO UI RNZ URIS SRN ANS RYO CRRA SRNR) ec Bi 3 RRR SOURS RESTS. 50 NASR RR CORR 9158 = ROU a RRR

oe -. ...LmUmr,rrw””:t~s””:”,”.”C~—rr—C—C“ERE FANON II Osi NOES ROR OSIGG NR CECH A OKC oo UN UG CCTM UNGER oS TARO cS ONG BCS NSIS EA Si

LO GC ee ee ae OCOe NG TRG SC URE oS TERS IG RI RRR 3 OR SO eNO NC eG

ASR TERR 3001/0070 5 RDI 0 0 RSS RE NO 2 RUNS ERR RB 2227 RR 7 TCO UU NSS GH TG

SEES IO INGSu es SIE IOS 5 RY RRS SD RS SMEAR Co UNRUH FURR oo) SOR ETO OST de RS RESTRE 5) URN SRN URS CSS3RRS NR 1EUG RRR RE US RH ROR NUS Sa Eel SA a So 1S Oe RURSSU SSUIAUS HEN ZS RUS RU CR NTIS OSC a

i MRI OIE Ota CON om EO

ol rrrrrCst«dsé RCs

BISON E NU RNG CONE EUR Shes URC NSS SUNG a AEE Ma NER ro SIN USURNG SS CONTEC a NUGENT UI INTC HAL REVA EEES ARC TULSA CHET NO UULUAET I ZEO RU RASC “CONCH OO NCNE

I CRON Se. OSES EES SE ISSN CI IU EGS Se RESON Ne INC BRI OO IE UUEE NS URS SIH NERC RGU Me NE Me aC ETL IC UST Te UR NIRUE ZA TNS COU ARNRSNC RNG RS TON

ARR URNS ONES NE RAE SURO SE GEE SESE USN ivatigatgn aie 551 ca MOMbaian SE? SOLOW NUMER MADSEN ip bp REG Eo eGR ERG EGON NS NSE nee Heo LOE MUU ORO NUS BOONES EZ SR RUSS EIGER! OS SEC AD EONS SSeS UAV RR CORN RUCK CEU) Mo TMM INCU US URS CEC ERR GURU CURT OAT

TEAR Cas TO GO COMM CB oe ee De Cf Re oe a OG Oe a an an PENSIONER U CSO ERA SRN LECCE SR ONCE BRS AB) UN ARR RUS 8 RR RRR ER SU 10 SE ROO OU NIN)

ee ee ee eeeeOe CS NN Se eo Re eG Gen Fe ee ee CO OOO a Gaa Oe ee OU Re Cali iia UN OSU STE eee COO ARS SOS ARETE AIS) RRR i I CEE NOONE SD ESSN GAR ROS BO CRRA Dy

DEVE ONS HONS Ae CRON RN NOR cae THE COCO RU RSE SLR SOS (715) RSE BRU UE PUES Oe oe GCN dE

Pe OE RSS OE OI ANG CNG SBC SO GSSiO ISG KS RM SCOURS oo i PR ON I SO NOES EU CO

tC ee a. EeooeeCeee7

ce INSSUH ROESGAC SUN DE aR2SAN ORM Suomi etseseT NS SO SOO GOCSS A CLO LODO WRC GLEN EO OCco NUE OU ONC GNIce CHO aoeONCE TEODOROOO NES ENO GAAMSEC NUR NTS ESO ANY GOURIRGS a fi ease deters EROS. FN ed ie CAME TSE RORY SACU SUITOR ES MLRe ALGORNATE US aa IMRAN SU EONAR AU CRBONS Reka: Geeta IRN iaia i cee ReHeRS FAG ORS a: PaesOUNEN 2 DISSE SESE Cea Ne ae ANSE BU SENG RO ONG

PUA UN CIT EON a OOM UU SL COU SUN AMCIS CUCU CGI ME RGR SUNG SO) SSTOORU a lec AlGe ean Me HAO EUG UN ZS aE ey

DOIN RAS OER RS OD ORE CC RRS RRC SONI SRR SM ORR UE RAIS eR SUNOS CRON US

TOSS OG a COVE SINR a SS OCC oN SE SNE 3 URE SoU oO EN ARTE EAN

ERO ONT RN eS SIN NER ENS NS RS SS OS ORCS RG IIRIR TORN Ge SR OK oS ero Ses WsROU 0 SU OaCSCIC. «NaoRSeaCOUT EN 0 BRU EOSIN CAISON OU RUHR VACUO Oal BUSOU OS SSS CG ON Ri oO) CUI ‘l CaaSONY a) BEE OBR RRR CBOE NS ANCOR INC GRE CUA CN

ee ee ee ee Ee Oa

OO CORR AC SURI EG 2009) 2 IS DORR RDS LH ee. ese Ma se oe ee Re Ot ama ce NOS OGLE ENS OURS UD

PROS ESE ROINIRCHT RUNS GLANS GUAM Ss EMS) a oe oo co ENERO ON BSNS SD HIS

CON An ae OR IGA OC OMNIS BRS IS CNMateOnRas NV ee ee SSUES ER SOE UAC SNR ASAIN SR LCE ONAL RSE NRE GE GSC SUSE OMINOUS UE EOE Sie UGH ESSN io rrr rtwsr—~—s:«Ci‘iCiéa‘izyCNSsrztzaSté‘CiéCOSCS;CCs#C.:Ci«sa‘aC(aiarss.tCi‘CiC;CCC

Le UIE ERO EGGS SSS OUI AG SUE UES EO Cs OR LOS GOI LCG OO OO ES UR OR UID RAO AUN Ua ARIGO TUNE DO IRONIC RAID AN RSIE OO SRURE RU BUN NOUNS ART EE SONI CEG Ge ENON UG HI AR Rent LEME CEN ALANA EES ODA OPS IRS NRT TUS tie NU NAO GG RS ITN ACERS URES ORO

POE CONES Soe GRU SOC HOSS ICRU SUS a NOC li SEs a EO Ge MORN GE SIU Se TNE ee OO AU UC

Fee OOeeaeine. =Ue|...a Oe ee BSNS OSU NEERING Us SNUARO ISIS SCOR Wan IO SHUTS UMUC Ress ES PSE OMG Os BECCA IVS Ve CGN ERT UR GLNTGIANG oe

OE MONG oo aeOSaee 4SISeas Luu Ree oe ee RO OO HUIS SAC NTRS eS RSC ORR SSO ONE RIA NSN IR— SSNS RN Uc HEU AROS SO UR QUE aN Pe MR IE a. CRIoe GO ee a aBE i Re Teme IN UCN UG ee RRO SEU NON HH RENO GU Us OUNG icG NU NGL SEU TIE EUG Oa BNE WU A GONE NS ANIA SUMO ER MNOS OU LOH SOG UNE TE ERS EG NAAT RU

ENR OS IE Cli SIMS 510° TA ARR CO) Sida acai itis Of much greater interest was the phenomenon of emperor worship and the extraordinary role of the slaves and ex-slaves in the imperial cults. The earliest of these, the Augustan

Lares, was in fact a revival of the dying Lares cult to which the emperor added his own imprint. Keith Hopkins argues that this cult had been started by ex-slaves, Augustus simply institutionalizing the informal local celebrations into a state cult devoted partly to his worship. “The cult provided rich ex-slaves, as organizers of the cult, with a prestigious and public outlet for social display. And it allowed emperor worship to flourish at street level.”'* It was not long, however, before emperor worship was accepted at all levels of society. It was a major legitimizing force among slaves for the simple reason that the emperor’s cult introduced into Roman law the alien principle of asylum for slaves. The granting of the nght of appeal to Caesar’s statue was

70 ‘The Internal Relations of Slavery one of the few ways in which the state intervened between master and slave. The state was, of course, sensitive to this intrusion on the authority of the master, and in practice very few slaves attempted such an appeal. But in enhancing the authority of the emperor in the eyes of all, including even the meanest of slaves, the legitimacy of the system as a whole was reinforced. What the master lost in individual authority, the slave system as a whole gained, embodied as it was in the divine protective power of the deified emperor.'”° Still, as Moses Finley has pointed out: “In so one-sided a relationship, in a world in which there was little hope of material success for the majority of the free population (let alone the slaves), and in which the earthly power was now pretty close to despotism, fear rather than love was often the dominating emotion behind worship, at best fear and love together. Religion became increasingly centered on salvation in the next world, whereas it had once been chiefly concerned with life in this one.”!*° Among the religions of salvation, Christianity was to emerge slowly, then dramatically, over the next three hundred years as the religion par excellence, one that could forge a moral order which appealed to and united

emperor and subject, master and slave.'’’ A discussion of the means by which it achieved this is beyond the scope of the present work. It is generally

accepted that Christianity found many of its earliest converts among the slave populations of the Roman Empire, although the fact is surprisingly difficult to authenticate.'** What is certain, however, is that the slave experience was a major source of the metaphors that informed the symbolic structure of Christianity.'”’

The most cursory examination of “the three terms which are the keywords” of the Apostle Paul’s theology (according to J. G. Davies) immediately reveals the extraordinary role of the slave experience as a metaphoric source. These key words are redemption, justification, and reconciliation.'”° Redemption quite literally means release from enslavement. Through Christ

the believer is emancipated from sin. Justification means that the believer has been judged and found not guilty, in much the same manner as the slave who has received the most perfect of manumissions, the restoration of his natality with the legal fiction that he had been wrongfully enslaved. “Reconciliation or Atonement means the bringing together of those who have been separated,” in much the same way that the manumitted slave is reborn as a member of a community. Paul in fact went so far as to use the idea of adop-

tion to describe the relationship between redeemed man and God. “Redeemed, justified, reconciled, man is elevated from the status of slave to that of son, and becomes ‘an heir through God’ of the promised salvation.”"*! What Ambrosio Donini calls “the myth of salvation” became the unify-

ing master concept of organized Christianity, and it is most powerfully evoked in the dominant symbol of the religion, that of the death and resurrection of Christ.'** Man fell into spiritual slavery because of his original sin.

Authority, Alienation, and Social Death 71 Slavery, which on the level of secular symbolism was social death, became on the level of sacred symbolism spiritual death. When, however, we question what Christ’s crucifixion meant, we find two fundamentally different symbolic interpretations. One explanation, which has profoundly conservative spiritual and social implications, held that Christ saved his followers by paying with his own life for the sin that led to their spiritual enslavement. The sinner, strictly speaking, was not emancipated, but died anew in Christ, who became his new master. Spiritual freedom was divine enslavement. Here was a confluence of two old ideas: the Near Eastern and Delphic notion of freedom through sale to a god, and the Judaic idea of the suffering servant and sacrificial lamb. It was not a very tidy symbolic statement, and it accounts in part for the occasional impenetrability of Paul’s theology. He had this interpretation in mind, for example, when he made remarks such as the following: “The death that he died, he died unto sin once: but the life that he liveth, he liveth unto God.” '* There was a far more satisfactory and at the same time more liberating symbolic interpretation of the crucifixion. The slave, it will be recalled, was someone who by choosing physical life had given up his freedom. Although he could, of course, have kept his freedom and died, man lacked the courage to make such a choice. Jesus, “his savior,” by his death made this choice for him. It is this feature that was completely new in the religious behavior and death of Jesus. What it meant in symbolic terms was that Jesus did not redeem mankind by making mankind his slave in the manner of the old pagan religions. Rather, he annulled the condition of slavery in which man existed by returning to the original point of enslavement and, on behalf of the sinner about to fall, gave his own life so that the sinner might live and be free. It is remarkable that Paul held also to this radically different interpreta-

tion of the crucifixion. The contradiction was directly paralleled by the well-known ethical contradiction of his theology. As Maurice Goguel has pointed out, Paul had two irreconcilable religious ethics.'** One was the pre-Christian and essentially Judaic ethic of law and judgment, in which obedience to divine law, and judgment according to one’s social and religious actions, were of the essence. The other was the ethic of the justified man. In this ethic Christ’s death redeemed mankind of the burden of sin; the believer, through faith, was immediately emancipated. The first ethic corre-

sponded to the conservative use of the slave metaphor; the second to the more liberal conception of slavery and emancipation. Paul tried to hold both these positions at the same time and thereby placed the believer, as Goguel points out, in the impossible position of one “who must struggle to realize in fact what he is in principle.”'*? And he asks: “How can we now speak of a

judgment for those who are in Christ, and therefore cannot be subject to condemnation?” !*°

The answer was to abandon the liberal view of emancipation and to

72 The Internal Relations of Slavery canonize the essentially pre-Christian interpretation of salvation as reenslavement to a god, in the triumph of the conception of the believer as the slave of God and of Christianity as a theological transmutation of the order of slavery. Whatever other factors explain Christianity’s conquest of the Roman world, there seems little doubt that the extraordinary way in which its dominant symbolic statements and meanings are informed by the experience of slavery was a major contributing factor. For the same reason too, Christianity was to provide institutional support and religious authority for

the advanced slave systems of medieval Europe and of the modern Americas. Christianity was not alone among the major world religions in legitimizing slavery. Earlier we noted the contradiction in Islam between the rationalization of slavery as a means of converting the unbeliever and the con-

tinuing enslavement of the converted. We find the same contradiction in Judaism and Christianity. The slave, in the city of the Christian God, was declared an insider, an integral part of the brotherhood of man in the service of God; but the slave, in the city of man, remained the archetypical outsider, the eternal enemy within, in a formalized state of marginality. At first sight the contradiction is not obvious. Indeed, the opposite seems to be the case: the exclusion of the slave on the secular level was symbolically compensated for by his inclusion in the sacred community. The contradiction between marginality and integration, which slavery created, was apparently resolved by relegating each to a separate domain of cultural existence. But this theological solution on the part of a monotheistic slaveholder class works only where there is hegemonic imposition of a rigid dual-

ism in the socioreligious ideology. This was exactly what happened in medieval Christendom under the conservative spell of Saint Augustine. '*’ But Judaism and Islam were too this-worldly and too strongly monistic for such an interpretation to be taken seriously.'** And the rise of Protestantism dealt a death blow to the neat symbolic compromise of the Middle Ages.

Augustinian dualism lingered in the symbolic representation of Latin American slavery: hence the apparent anomaly that has baffled so many Anglo-American historians, that of a Catholic church stoutly declaring slavery a sin, yet condoning the institution to the point where it was itself among the largest of slaveholders.'*° The Anglican masters of the Caribbean avoided the problem altogether by abandoning religion or making a mockery of it, both for themselves and for their slaves—clergymen in nineteenth-century Jamaica being “the most

finished debauchers in the land.”'*? As Richard S. Dunn has pointed out, the refusal of the English planters in the West Indies to convert their slaves to Christianity, in contrast with contemporary Latin masters, “can largely be explained by Protestant versus Catholic conversion techniques.”!! Protestantism by its very nature demanded the liberating conception of the cruci-

Authority, Alienation, and Social Death 73 fixion, with its emphasis on personal choice and freedom. Realizing this, the

West Indian masters did everything possible to keep their slaves in ignorance of their creed—giving in only when, a few decades prior to abolition, they found their policy to be too easy a target in the propaganda war of the abolitionists.

How then do we account for the Protestant slave South where, during the late eighteenth and the nineteenth centuries, both masters and slaves were highly religious? It is clear that the special version of Protestantism that triumphed in the South and the peculiar socioeconomic features of the system together explain its unusual course of development. Until nearly the end of the eighteenth century the U.S. South did not dif-

fer markedly from other Protestant slave systems. Masters were generally hostile to the conversion of their slaves, fearing—like their Caribbean counterparts—that the nature of their creed with its emphasis on instruction in the gospels, personal choice, and spiritual liberation would, if adopted by their slaves, undermine the masters’ authority. As late as 1782 slaves in Georgia were still being whipped savagely for preaching,'”* and while Albert Raboteau may have overstated the case in claiming that “the majority of slaves... remained only minimally touched by Christianity by the second decade of the nineteenth century,” he was not far wrong.'”° Two major developments explain the remarkable change that took place during the nineteenth century. One was the great religious awakening that culminated in the religious conversion of the South from classical Protestantism to revivalist fundamentalism.'’* The second was the emergence in the South between 1790 and 1830 of a full-fledged slave system, a total commitment to the institution as an essential feature of the region’s socioeconomic order, and the realization that if slavery was to function effectively the system had to be reformed. In Genovese’s words, “whereas previously many slaveholders had feared slaves with religion, now they feared slaves without religion even more. They came to see Christianity primarily as a means of social control ... The religious history of the period formed part of the great thrust to reform slavery as a way of life and to make it bearable for slaves.”!”° Fundamentalist. Protestantism was peculiarly suited to such a reform. Its emphasis on conversion as a sudden spiritual transformation rather than the

result of reflection and instruction; its oral rather than literary missionary techniques; its other-worldliness, especially its insistence on salvation as a purely spiritual change, the rewards of which are to be achieved in the hereafter; its emphasis on piety and obedience, and on the sinfulness of the world and the flesh; made it a creed that the masters could confidently regard as a support for, rather than a subversion of, their authority.'”° Nevertheless, it would be simplistic to interpret the role of religion in the slave South solely in terms of an opiate for the masses, a device used by

74 The Internal Relations of Slavery the master class as an agent of social control. In the final analysis it was indeed just that, and there is abundant evidence that the master class cynically devised a “theology of slavery” in a crude attempt to rationalize the system. But as recent studies have shown, slaves quickly recognized the crude ideological strategy of their masters. Olli Alho’s detailed analyses of the slave narratives “indicate that the carefully constructed theology of slavery built

up by the whites became in many plantations nothing more than a joke” among the slaves.'”’ The slaves found in fundamentalist Christianity paths to the satisfaction of their own needs, creating the strong commitment to Christianity that has persisted to this day. In so doing they created an institutional base that pro-

vided release and relief from the agonies of thralldom, and even offered some room for a sense of dignity before God and before each other. Having said all this, | must emphasize that the religion they experienced was the same as that of their masters in all its essential doctrinal and cultic aspects; that while the spirituals they sang may have had a double meaning with secular implications, it is grossly distorting of the historical facts to claim that they were covertly revolutionary in their intent; and, most important of all, it

is irresponsible to deny that however well religion may have served the slaves, in the final analysis it did entail a form of accommodation to the system.

In all of this I am in complete agreement with Genovese’s penetrating interpretation of the role of religion in the slave South.'°* Where I differ from him, and from others such as Lawrence W. Levine’”’ and Albert J. Raboteau who with equal skill and persuasion have emphasized the creative and positive side of religion for the slave, is in my interpretation of the specific means by which fundamentalist Christianity became at one and the same time a spiritual and social salvation for the slaves and an institutional support for the order of slavery. To appreciate where we differ it is necessary to return to the nature of Christianity and to specify the peculiar doctrinal features of fundamentalism. Pauline Christianity, as we saw, was theologically dualistic, containing an ethic of judgment and an ethic of the justified person that were in constant tension with each other. These two ethics in turn were symbolically expressed in two contrasting interpretations of Jesus’ crucifixion. Roman Catholicism resolved the tension by eliminating what I call the liberating pole of Pauline dualism, emphasizing the ethic of judgment and obedience; classic Protestantism resolved it by eliminating the conservative pole and by strongly reviving the ethic of the justified person.

What then is distinctive in fundamentalism? My answer is that it restored both poles and returned fully to Pauline dualism with all its contained tension and its contextual shifting from one ethical and symbolic pole to the

other. If we do not understand this distinctive doctrinal feature of fundamentalism, we cannot fully appreciate how the religion could have spirit-

Authority, Alienation, and Social Death = 75 ually sustained both slaves and masters as well as the system as a whole. We will also fail to comprehend the symbolic life of the slaves themselves. If we next seek the major doctrinal and symbolic components of slave religion, we find that the fundamentalism of the slaves was, like that of all

southerners, essentially Pauline in its overwhelming preoccupation with Christ and the crucifixion and in its ethical and symbolic dualism, its paradoxical tension between the ethic of judgment and the ethic of the redeemed sinner. Further, it is precisely this dualism that explains the apparent paradox that the religion of the slaves, doctrinally one with that of their masters, nonetheless allowed for the spiritual support of both groups and of the sys-

tem as a whole.

Jesus and his crucifixion dominate the theology of the slaves and not, as recent scholars have claimed, the Israelites and Exodus story.'®’ Not only is the theme of the crucified Christ explicitly central and dominant, but even

when figures from the Old Testament are referred to (including Moses), closer examination reveals that the allusion is really to Jesus. Although Alho does not make the connection to Pauline theology, it is striking that his most important finding concerns the dualistic conception of Jesus in the religion of the slaves—that of Jesus as Messiah King and Jesus as comforting savior. He concludes his interpretation with a reference to an insightful contemporary observer: “The difference between the two main identities of Jesus reminds one of what T. W. Higginson wrote in his camp diary about the religious behavior of his black soldiers; softness, patience, and meekness on the one hand, hardness, energy, and daring on the other, seems to be reflected in

the dualistic way in which the spirituals picture the figure and roles of Jesus.”'°! We can now explain how fundamentalism, a single religion, performed the contradictory roles it did in the slave South. Both masters and slaves ad-

hered to Pauline ethical dualism, with its sustained “eschatological dissonance.” '®* And in exactly the same way that Paul and the early Christians shifted from one pole of their doctrinal dualism to another as occasion and context demanded, so did the masters and their slaves. Thus the masters, among themselves, could find both spiritual and personal dignity and salvation in the ethic of the justified and redeemed sinner. The crucified Jesus as redeemer and liberator from enslavement to sin supported a proud, free group of people with a highly developed sense of their own dignity and worth. Similarly, the slaves in the silence of their souls and among themselves with their own preachers, could find salvation and dignity in this same interpretation of the crucified Lord. When the theologian Olin P. Moyd insists that “redemption is the root and core motif of black theology” and that it means essentially liberation from sin and confederation within the fellowship of black worshippers, it is, I suspect, to this end of the Pauline dualism that he is referring.'® As with the masters, the slave dualism had another pole. This is the ethic

76 The Internal Relations of Slavery of law, judgment, and obedience, the ethic that found symbolic expression in the other Jesus, the more Judaic Messiah King who judges, who demands obedience, and who punishes the wicked and rewards the righteous. This is the Jesus who saves not by annulling slavery but by divine enslavement. To live with this Jesus demands, as Goguel tells us, watchfulness, obedience, and stoic acceptance. Both masters and slaves held also to this conception of Jesus and, like Paul and the early Christians, shifted to this symbolic code in dealing with, and coming to terms with, all authority relations—not only the one between master and slave but, among the masters between male and female, upper class and working class, parent and child, and among the slaves between parent and child. In this way fundamentalism, by reverting to Pauline dualism, provided the slave South with the perfect creed, one much more subtle in its support for the system than most of the masters thought. The crude theology of slavery that the masters tried unsuccessfully to preach in the plantation mission was really quite unnecessary. Nor was it necessary for master and slave to have two separate religions. Christianity, after Paul, had already constructed an extraordinarily shrewd creed with a built-in flexibility that made it possible for emperor and slave to worship the same god without threatening the system, but also without denying all dignity to the oppressed. In the U.S. South there developed the last and most perfectly articulated

slave culture since the fall of the Roman Empire. The religion that had begun in and was fashioned by the Roman slave order was to play the iden-

tical role eighteen hundred years later in the slave system that was to be Rome’s closest cultural counterpart in the modern world. History did not repeat itself; it merely lingered.

Honor

and

Desradati esradation

NEAR THE MIDDLE of the first century B.c., the mimewriter and epigrammatist Publilius Syrus triumphed over his rival Laberius in a dramatic contest of verbal skill ordered by Caesar. A major factor con-

tributing to his success must certainly have been what J. Wight Duff and Arnold M. Duff described as his “gift of understanding Roman psychology.”' Publilius’ comprehension of the Roman mind came from a very special perspective, that of the underdog—that sharp, incisive insight and sardonic wit honed on bitter experience which gives the most trusted and intelligent members of an oppressed class a distinctive access to the mind of

their oppressors. For Publilius had been a slave of Syrian origin, brought possibly from Antioch as a youth, who by sheer force of intellect and verbal skill in the language of his master won both his freedom and the adoration of the populace. As we examine the remnants of the maxims extracted from his mimes, it comes as no surprise that a disproportionate number of them are concerned with the nature of honor and the indignities of submission. When he wrote that “the height of misery is to live at another’s will,” the words came from deep in the suffering of his own earlier life.” Publilius knew how to play on the deepest weaknesses, anxieties, and conceits of the various categories of persons in his audience. He no doubt had the slaveholders in mind when he inserted in one of his mimes, “Honor 77

78 The Internal Relations of Slavery scarce ever revisits the mind it has quitted”;’ and it was to placate the freedmen that he threw in, “None ever loses honor save him who has it not.’’* But I strongly suspect that it was for the slaves who looked on from the fringes that he coined his finest maxim: “What is left when honor is lost?””? There was no need to elaborate, for everyone——master, freedman, and most of all slave—-at once knew the answer. And so would the members of all other societies in all other times. The idea that a person’s honor is more valuable than his life, and that to prefer life to honor betrays a degraded mind, comes close to being a genuinely universal belief. It is a theme that haunts Western literature. Pascal need not have been influenced by Publilius when over fifteen hundred years later he expressed the view that “he would be infamous who would not die to preserve his honor.” The most cursory search will quickly turn up numerous

similar passages, from Shakespeare’s Richard the Second proclaiming, “Take honor from me and my life is done,” to Nietzsche’s superman declaring, “One should die proudly when it is no longer possible to live proudly.” And even where writing is not to be found, the proverbs and oral traditions of all preliterate peoples, including every headhunting and cannibal tribe I know, would reveal an almost identical belief: that to choose life over honor is infamy. As Marcel Mauss has observed: “Even in really primitive societies like the Australian the ‘point of honour’ is as ticklish as it is in ours... Men could pledge their honour long before they could sign their names.’”®

Yet it was the choice of life over honor that the slave or his ancestor made, or had made for him. The dishonor of slavery, I have already argued, was not a specific but a generalized condition. It came in the primal act of submission. It was the most immediate human expression of the inability to defend oneself or to secure one’s livelihood. It was not part of the institutionalization of slavery, for its source was not culture. The dishonor the slave was compelled to experience sprang instead from that raw, human sense of debasement inherent in having no being except as an expression of another’s being. What the captive or condemned person lost was the master’s gain. The real sweetness of mastery for the slaveholder lay not immediately in profit, but in the lightening of the soul that comes with the realization that at one’s feet is another human creature who lives and breathes only for one’s self, as a surrogate for one’s power, as a living embodiment of one’s manhood and honor. Every slavemaster must, in his heart of hearts, have agreed with

Nietzsche’s celebrated declaration: “What is good? Everything that heightens the feeling of power in man, the will to power, power itself. What is bad? Everything that is born of weakness. What is happiness? The feeling that power is growing, that resistance is overcome.”’ We are far removed from the ostentatious ironies of Nietzsche when we turn to the essays of Francis Bacon; but in his own quiet, almost serenely

Honor and Degradation 79 self-assured way, for his own time, Bacon made the same point when he wrote, “Discreet followers and servants help much to reputation. Omnis fama a domesticiis exeant [All reputation proceeds from servants].”* A story from the Icelandic sagas provides another illustration. It concerns the behavior of the crafty slave overseer Atli, who without permission lavishly entertained the shipwrecked Vebjorn and his men throughout the winter at his master’s expense. When the master, Geirmund, finally learned of his slave’s extravagance, he was understandably outraged and demanded an explana-

tion. The cunning Atli responded that “he wanted to show how greatminded and generous his master must be, and how great an estate he must have, when one of his thralls dared to do such a thing without his permission. Geirmund was so pleased with the answer that he gave the thrall his freedom and also the farm he had managed.”

The Nature of Honor In this chapter I propose to show that, first, in all slave societies the slave was

considered a degraded person; second, the honor of the master was enhanced by the subjection of his slave; and third, wherever slavery became structurally very important, the whole tone of the slaveholders’ culture tended to be highly honorific. (In many societies the sole reason for keeping slaves was in fact their honorific value.) Before considering the comparative data, we need to clarify the concept of honor by drawing on the rich anthropological literature on the subject. Happily, a communis opinio has emerged and is well expressed in the work of Julian Pitt-Rivers.'° He argues that honor is a complex notion having several facets: “It is a sentiment, a manifestation of this sentiment in conduct, and the evaluation of this conduct by others, that is to say, reputation. It is both internal to the individual and external to him—a matter of his feelings, his behavior, and the respect he receives.” The way in which these facets of behavior are related is cogently put: “Honor felt becomes honor claimed, and honor claimed becomes honor paid.” Only those who aspire to honor can be dishonored: “Those who aspire to no honor cannot be humiliated.”

What this immediately implies is that those who do not compete for honor, or are not expected to do so, are in a real sense outside the social order. To belong to a community is to have a sense of one’s position among one’s fellow members, to feel the need to assert and defend that position,

and to feel a sense of satisfaction if that claimed position is accepted by others and a sense of shame if it is rejected. It is also to feel that one has a right to take pride in past and current successes of the group, and to feel shame and dishonor in its past and present failures. Modern anthropologists have confirmed Thomas Hobbes’ insight that the sense of honor is intimately related to power, for in competing for prece-

80 The Internal Relations of Slavery dence one needs power to defend one’s honor. Still, it is one’s sense of honor that often drives one to acquire the instruments of power in the first place. It is foolish to insist too strongly on a single causal direction, and even worse to

adopt a vulgar materialistic interpretation of the relationship. It is true, as John Davis observes, that honor “describes the distribution of wealth in a social idiom, and prescribes appropriate behaviour for people at various points in the hierarchy”; in this sense “it entails acceptance of superordination and subordination.” But, he goes on to note, honor “1s closely associated with integrity: the whole man is contemplated.”’"’ Pitt-Rivers finds that “the claim to honor depends always in the last resort, upon the ability of the claimant to impose himself. Might is the basis of right to precedence, which goes to the man who is bold enough to enforce his claim, regardless of what may be thought of his merits.” It 1s for this reason that “courage is the sine qua non of honor, and cowardice its converse.”'” Because honor envelops “the whole man,” it is seen as an intimate personal quality relating to both his physical and characterologic attributes. A person’s will and intentions are the two vital ingredients in any assessment of his honor by others. Is he a man of his word? Is his oath inviolable? Can he assert his will as a man of honor? “The essence of honor is personal au-

tonomy,” and to be in another’s command, Pitt-Rivers adds, “restricts it.” Furthermore, a freely established relation of dependence with a more powerful patron can be the basis for expanding one’s honorific claims vis-a-vis one’s equals. The client’s attachment also firmly establishes him in a place within the hierarchy of honorable statuses. He belongs and is one with his patron as a member of their society. The patron needs him as much as he needs the patron, and this is fully understood by both parties. The idea that honor is personal autonomy takes us to the philosophical core of this most elusive of social concepts. For the real mystery of honor lies

in the fact that although its existence is revealed, and its claims proven in acts of honor, such acts are always considered epiphenomenal. This should be evident from the common observation that two persons may perform the same act, yet the behavior of one is considered honorable while that of the other is not. Acting honorably is not the same thing as being honorable; it is not enough to abide by a code of honor. Honor is never evaluated in teleological terms. Like Immanuel Kant’s “good” which is nothing if not a “good will,”'’ honor is nothing except the honorable will. Nor does the fact that one is honored make one honorable. One need not even be a human being to be honored; in India millions of persons daily honor the cow. Understanding this aspect of honor is critical, if what I have to say later is not to be

misunderstood. There have been slaves who have been honored or whose acts have been considered honorable, yet who have remained despised as persons without honor. Finally, we must take account of the important role of honor as a dis-

Honor and Degradation 81 tinctive feature of certain cultures. The sense of honor is present in all human societies—in some to the point where it becomes a dominant value. Following Plato, we may call the culture of such societies, and the character syndrome in which honor and pride are excessively developed, timocratic.'“

Honor and Slavery among Tribal Peoples Let us begin with the Tupinamba of South America, a primitive, warlike group among whom slavery existed in its most elementary form. Economic motives were wholly absent in the enslavement of captives. Slaves were kept for two purposes only: as a living exhibition of the master’s honor and valor

in war, and ultimately as meat for the cannibalistic orgy that might take place as long as fifteen years after capture. Between being taken prisoner and being eaten, the captive “recognize[d] himself as a slave and a defeated man, he follow[ed] the victorious man, serve[d] him faithfully without having to be watched.””° The slave among the Tupinamba was constantly aware of the fact that he was a doomed person. Even if he escaped, his own tribe would not take him back. His sense of degradation was as intense as his master’s sense of

glory. A Tupinamban slave told Father Evreux that what really bothered him was not the prospect of being eaten, but not to be able to take revenge before dying on those who are to eat me. I

remember that I am the son of an important man in my country ... Now I see myself as a slave without being painted and no feathers attached to my head, my arms, around my waist, as the important people of my country are decorated, then I want to be dead.'®

While slaves among the primitive Germanic peoples may not have been

physically consumed, it is no exaggeration to say that they were socially consumed (as they were to be in all other slaveholding societies). Carl O. Williams’ observation on slavery in ancient Iceland is pertinent: The class of the lowly is the source from which the master class draws its livelihood and leisure. Thraldom is a degree of cannibalism. It is a system of man feeding upon man. The master is a human parasite, who, by the right of might, has secured his fellow-men in the bonds of thraldom in order to feed upon them and to use them for the satisfaction of his appetites.'’

What the slave mainly fed was the master’s sense of honor and his sexual appetite, for the economic role of the slave was quite marginal among most

of the continental Germanic tribes.'* Among these peoples, however, the sense of honor was highly developed, each nonslave member of the community having a specific honor price determined by his position in the kin group and the group’s position in the wider community. Slaves were regarded, above all, as people without honor. This could take rather amusing

82 The Internal Relations of Slavery turns. For instance, if a member of the community verbally abused another he could expect savage retaliation, sometimes resulting in death; slaves, however, could verbally abuse anyone if they were so inclined, because “the abusive language of a slave cannot injure anybody’s honor. If his abuses become offensive, the slave must be looked on only as the mouthpiece of the lord.”!” Of course, if the slave made a nuisance of himself he could be killed on the spot, but the matter of honor was irrelevant. If the freeman chose to laugh the matter off as the simple rantings of a crazy brute, there was no loss of honor. This was true of all the Germanic tribes, with minor variations. Much the same situation prevailed among the Norwegians and their Icelandic offshoot. “No one,” declared the Icelander Hovamol, “should put faith in a sick calf or in a self-willed thrall.”*° It was a mortal insult among the Icelanders to call someone a thrall, for it amounted to saying that they were without honor. To do so was a fullrettisord (a gross verbal insult requiring atonement) and invariably resulted in bloodshed.*! As among other Germanic peoples, the injury or murder of a thrall required compensation to the master, but the compensation was in no way viewed as part of an honor price payment. The master might, if so inclined, view the injury as an offense to his own honor, but even then “the offense was of no grave consequence—only a matter of a boot of twelve aurar [the value of twelve cows]

for each thrall.”

The same situation existed among the Welsh and Anglo-Saxons. To be sure, the Welsh laws required a sarhed, or honor payment, for injuries to slaves; but a closer examination of the laws reveals that the payment was to be made in kind not to the slave but to his master, and the goods specified all related to materials for the improvement of the slave’s working capacity. Even where a slave woman was sexually abused, the sarhed was to be paid not to her or her common-law slave spouse, but to the master.”? A similar situation existed among the Anglo-Saxons during the seventh century. The honor price for raping a “birele” or household slave, H. R. P. Finberg tells us, was “appointed in proportion not to her feelings but to her master’s rank: 12 shillings or 240p for a nobleman’s, 6 shillings or 120p for a commoner’s” (emphasis added) and so on down to the “twenty-five shillings women” or “srinding slave.” The money apparently went to the master.” If we move now to the impressive body of data on domestic slavery in sub-Saharan Africa,”° there is unambiguous support for our argument. All traditional African societies were extremely alive to the role of honor in people’s lives. Where large-scale slave systems existed, honor became a

dominant value. The classic examples were the Nigerian pre-European states such as Bornu and Hausaland, the Amhara of Ethiopia, and the nineteenth-century Ashanti.”° In the great majority of traditional African societies, however, stratificatory systems were not highly developed and classes

Honor and Degradation 83 were either absent or not “well defined.””’ However, precisely because such classes and status groups were not well developed, individual competition for honor and prestige was rampant—as is well known to students of preindustrial societies.** The less centralized were such societies, the greater their emphasis on prestige ranking of individuals. (The extreme was the highly

formalized recognition of honor found among the largely acephalous Ibos.*’)

In the struggle for prestige, what was critical in al] African societies was

the number of dependents an ambitious man could acquire. Kinship and affinal alliances were the two major techniques for accumulating dependents, but a third important means was the institution of slavery. Among many African tribes this was often the sole reason for the acquisition of Slaves, there being little or no economic difference between the condition of slaves and their masters and no such thing as a slave class.*° Typical of the African situation were the Mende, who kept considerable numbers of slaves for both social and economic reasons and on the whole treated them well, so well indeed that it was difficult for an outsider to distinguish between free and slave. The primary social difference between the two groups was the honorlessness of the slaves, a condition that the free man was reluctant to point out in the presence of strangers, knowing how crushing it would be to the slave. Thus J. J. Grace tells us that the nduwanga or

slave group was subjected to “a prescribed code of conduct which made their inferior status clear.” He cites T. S. Allridge, a former trader and official in Mende country, who wrote: “Slaves merely cringe up and place their hands one on each side of their master’s hand, and draw them back slowly without the fillip while the head is bowed.””*' The loss of honor was most evident among aged slaves. In no other part of the world was age more respected and honored than in traditional African societies. But old Mende slaves never received this respect: “They were minors who would never re-

ceive the respect due to a mature adult.” Nowhere in Africa was the association between slavery, the timocratic character, and the conception of the slave as a degraded person more pronounced than in the large-scale slave societies of the Fulani. In his brilliant study of Fulani society in Jelgobi (on the Upper Volta), Paul Riesman shows how the strong image the Fulani have of themselves is negatively defined largely in relation to their stereotype of the despised maccube (slaves) and ex-slaves. In Fulani eyes, it is among “captives” or ex-slaves that one finds most clearly expressed everything that is the opposite of Fulani. According to this stereo-

type, “captives” are black, fat, coarse, naive, irresponsible, uncultivated, shameless, dominated by their needs and emotions. These qualities are innate and manifest the servile condition, for the Fulani cannot imagine that a descendant of slaves could have better qualities than his ancestors.”

84 The Internal Relations of Slavery The term pulaaku means everything that is ideally Fulani, and Riesman found that the best way to define it was simply to “make a list of antonyms of the terms which define the stereotype of the maccudo. It follows that the Fulani should be: light-skinned, slender, refined, subtle, responsible, cultivated, endowed with a sense of shame, and master of his needs and of his emotions.” The Fulani ideal is strongly expressed in Fulani epic poetry, in which the “very word pulaaku has a meaning which obliges us to put the accent on the social: pulaaku means not only ‘the qualities appropriate to a Fulani’ but also at the same time the group of Fulani men possessing these qualities.” In other words, there is not only a timocratic character, but a timocratic group and culture. Significantly, Riesman finds that the Somali

bard’s use of the term pulaaku “is an exact structural equivalent of the English word ‘chivalry’ and, like it, designates at once certain moral qualities and a group of men possessing these qualities.”** Among the Indians of North America hereditary slavery of any significance existed, with only a few exceptions, mainly on the northwest coast. There the condition of the slave was unenviable—but rarely for economic

reasons, since the considerable surplus generated from rich fishing beds meant that the consumption patterns of masters and slaves were much the same. Slaves, however, were utterly without power or honor. What Robert E. Stearns wrote of the Kassi tribe in the late nineteenth century was true of the entire northwest coast during this period, that “they treat their slaves as if they were dogs; they look on them as a possession outside the human category. For a master to kill a dozen slaves is nothing; it merely demonstrates his wealth and his power.”* In the potlatch ceremony, for which these peoples are best known, the killing of slaves often reached frightful proportions, especially among the Tlingit. There is no reason for us to become entangled here in the vexatious problem of the potlatch, a subject on which American anthropologists have been waging intellectual war since the turn of the century.*’ Incontestably though, the ceremony was closely tied to the Indians’ excessively developed sense of prestige and honor (whatever other functions it may have served), and the slaughter, freeing, or donation of slaves was its high point. Although slaves served some economic functions, as in Africa and South America, their primary function was to support the honor and power of their masters. Perhaps the most remarkable instance of the association between slavery

and both honor and degradation among preliterate tribal peoples is the Toradja tribes of the central Celebes. This group of tribes is instructive for two reasons: first, some had a highly developed system of slavery with a large slave class, while others did not, hence valuable comparisons can be made; and, second, the relationship between slavery and culture was direct. The first lesson we draw is that among the tribes with a slave standing

Honor and Degradation 85 work had become dishonorable.*® Second, while decisions were made in a democratic consensual manner among the tribes that kept no slaves, deci-

sion making among the slaveholding tribes was highly autocratic: “The manner in which a Lage chief handles matters testifies to a feeling of power that has developed through mastery over his slaves, but from which the free

in the society also feel the influence.” Third, childrearing patterns were strongly influenced by the large number of slaves: children in the slaveholding groups were far more obedient and grew up to be far more authoritarian than did those of the nonslaveholding tribes. Fourth, the sexual exploitation

of slave women and the resulting “licentiousness of many female slaves made the free women more prudish.”” What resulted from all this was an almost perfect example of the timocratic character among the slaveholders, in sharp contrast to the highly sensitive, give-and-take attitude of members of nonslaveholding tribes. According to N. Adriani and Albert C. Kruyt: Slavery has put its stamp on the character of the various Toradja tribes. The To Lage and the To Anda’e, who always had to be mindful of keeping their prestige high with regard to their slaves, had in this way achieved a great deal of self-control, through which they made a more civilized impression on the foreigner than did the To Pebato who, not knowing this pressure, behaved more as they are, let themselves go more . .. The feeling of responsibility that is characteristic of the To Lage must also be a consequence of keeping slaves, since the lord answered for his slaves and was responsible for

their deeds.*” ,

At the same time, the personality of the slave was considered to be totally opposite from that of the master, and the slave’s behavior tended to conform to the master’s view of him. How he actually felt, of course, the ethnographers do not tell us, but their description of his social character is typical of the political psychology of slavery in all times and all places: There was indeed a great difference in character and disposition between the free and the slaves. The slave is so accustomed to not being allowed to have any free will that he has a great deal of passiveness and indifference. One therefore cannot depend on him. The slave has little feeling of responsibility for-his deeds; if he has done something wrong, his master is there to pay the fine for him. In the rice field or at the salt-making place, he is equally indolent everywhere, because he knows that he does not work for himself, and only the presence of his master will move him to moderate effort. Because he is not heard at deliberations over political or social matters, he is dull and indifferent about everything that happens in the village and tribe. Because of all this slaves are often rude, and aba mbatoea, “slave manners,” is tantamount to improper behavior.*’

86 The Internal Relations of Slavery Honor and Slavery among Advanced Premodern Peoples The situation was much the same among the slaveholding peoples of the advanced civilizations, although there was not always the same direct causal link between slavery and the development of timocratic character and culture.

Throughout all the advanced oriental societies the sense of honor was heavily stressed.** It was ideologically elaborated among the Chinese and the many peoples they influenced in the distinction between base and ignoble or dishonorable persons. All imprisoned criminals lost honorable status and were “base,” and E. G. Pulleyblank propounds the generally accepted view that “it was at least partly a result of the penal origin of slavery that slaves were termed chien (‘base,’ ‘ignoble’) as opposed to the normal population who were called liang, ‘good.’ 43 Criminals, while assimilated to the status of slaves, were not all slaves; for the condition of some of them was

only temporary. Furthermore, it was always illegal in China to reduce a “good” or “honorable” person to slavery, although this dictate did create some rather tricky legal problems. The same was true of Korea, where “the good people consisted of Yangban [aristocrats] and commoners,” while the

“base” consisted “primarily of slaves” (although certain polluting caste members were also base).** The Gia-Long code of the ancient Vietnamese spelled out this distinction in fine detail as it applied to persons in genuine slavery as opposed to those in debt-servitude. Hired or pledged persons, the law stated, “are on the same footing as persons of honorable status and are considered as ‘any persons whomsoever’; one may not consider them equivalent to those who are slaves in perpetuity.”*° In relation to their owners and the families of their owners, pledged persons and others in debt-bondage had little power and were totally dependent economically. In that respect their condition resembled slavery, but it would be a great mistake to confuse them with slaves, for they were not without honor. Dang Trinh Ky comments, “In social life, in relations with strangers, they are always considered as ‘any persons whomsoever’ and ‘honorable.’ They have never lost their

dignity or liberty.“ In ancient India we find the same conception of the slave as a “base” person, which is not surprising in view of the Buddhist influence on this mode of conceptualizing the absence of honor.*’ While not conceptualized in the same way, the base condition of the slave was even more pronounced in the pre-Buddhist era. In the Rigveda the slave is not even considered a human being.** During all the periods of ancient India the term “dasa,” meaning slave, was always a term of abuse, and according to Katualya it was a crime to call someone “dasa” or “dasi.”*° In considering the ancient Greeks, we encounter not only one of the two most advanced slave systems of antiquity but, not accidentally, a society in

Honor and Degradation 87 which, on the one hand, the degraded condition of the slave was consciously articulated and, on the other hand, the culture was highly timocratic. I am certainly not suggesting that large-scale slavery was the cause of the honorific nature of classical Greek culture. The world of Odysseus, as Finley has shown, had an extremely honorific culture with slavery being of only marginal significance.” What I do maintain is that in classical Greece, slavery

and the timocratic character were mutually reinforcive. The preexisting timocratic value system, along with new economic forces, encouraged the development of large-scale slavery. At the same time, the enormous growth of slavery not only reinforced the timocratic character of the ruling class but stimulated its diffusion among all classes, for by the classical period these were societies in which even the destitute felt deprived if they could not af-

ford a slave.”'

Few would disagree with Alvin Gouldner when he says that “a central, culturally approved value of Greek life, embedded in and influencing its systems of stratification is an emphasis on individual fame and honor,” and that the contest for power and honor in ancient Greece, as in most honorific cultures, was largely a zero-sum game, “in that someone can win only if someone else loses.”””*

It is not difficult to imagine the extraeconomic role of slaves in such a society. After reviewing the intellectual evidence, Robert Schlaifer sums up

the popular conception of the slave as one who was “completely without honor, shame or any sound element at all.””°’ The slave was a stock character

in Greek comedy and even Joseph Vogt, after reviewing the literary evidence, was obliged to conclude that “slaves were irrevocably degraded in the eyes of the public.”°* Where the slave was not cast in his usual role as a lazy, cowardly buffoon, it is significant that he was merely an onlooker to tragedy. He himself never experienced tragedy, and was never “allowed to participate in anything even remotely connected with suffering or responsibility.””° While it is true that the law of hubris held it an offense to outrageously insult a slave, G. R. Morrow goes too far in claiming that it implied either respect or unusual protection for the slave.*° As he himself observes, fourth-century orators found it anomalous that this law should apply to slaves, “for what honor has a slave to lose?””’ The law during the classical period must cer-

tainly have been a dead letter, since the court was made up of the citizen body and its verdicts strongly reflected public opinion. What is more, the slave had no legal standing and could not bring charges in this court. It is difficult to conceive of a situation in which a third party would bring charges on a slave’s behalf against a free person, in view of the fact that the accuser faced a stiff penalty if less than a fifth of the court supported the charge.”

There was little humanity in the conception of the slave in classical Greece, and Finley has effectively demolished attempts by some classicists to suggest that the slave was treated as something more than an utterly de-

88 The Internal Relations of Slavery graded figure.”’ “Both male and female slave were an ‘unfree body,’ andra-

poda, ‘human-footed stock,’”°’ and like their Mende counterparts, old slaves could anticipate no respect: “One of the favorite etymological jokes

often.’”°! Se , ,

was to derive the word for ‘boy’ and ‘slave’ from the word for ‘to strike’; thus

even an old slave could be addressed as ‘boy’ because he was beaten so While it is easy to show the existence side by side of a highly timocratic

elite culture and a degraded slave condition, it is quite another matter to demonstrate how the two related to each other. Indeed, all we can do is speculate on the nature of the relationship. The best guess is that the large-scale slavery and the timocratic culture of classical Greece had independent historical sources before they came to reinforce each other. The contempt for working for others—and among the ruling class, working at all—must certainly have encouraged the growth of slavery. Further, it is not unreasonable to speculate that slaves did more than help in meeting material needs, they

also satisfied a psychological need to dominate. As Victor Ehrenberg observes: “Free men and women frequently indulged their pride towards slaves without restraint. The master was always the absolute lord and owner, the despotes.”® Slaves, moreover, may well have had a direct effect on the character for-

mation of the Greek middle and upper classes, in view of their important role in bringing up the children of their masters. But Finley has, quite reasonably, expressed reservations about drawing conclusions from this function about the attitudes of masters and slaves toward each other.” I fully agree that the use of male and female governesses and nursemaids in no way

encourages “humanity” on the part of a master class toward the class of people who rear them; the experience of black southerners in the United States should have made that clear to even the most anticomparativist of classicists. At the same time the southern experience suggests that a dependence on slaves for childrearing does have some effect on the character formation of the children involved; that, in short, it reinforces arrogance and

authoritarianism and supports the timocratic syndrome.” When, further, one takes account of the peculiar status of women in classical Greek culture, the ready availability of female slaves as sex objects, and the apparent tendency of the father to absent himself as much as possible from the home,” it

appears that among upper-class Greeks of this era the role of household slaves in the formation of the timocratic character was not unimportant.

We now move to a consideration of the ancient Roman experience, which for four reasons is critical to my argument. First, Rome evolved the most complex slave system of all the peoples of the premodern world. Second, the Romans, like the Greeks, had a remarkably developed sense of honor. Third, Rome presents the unusual, though by no means unique, case in which significant segments of the slave population—the Greeks and other

Honor and Degradation 89 hellenized slaves—were acknowledged by their masters as culturally superior. Finally, in Rome a group of slaves and freedmen exercised extraordinary power in both the executive and administrative branches of the imperial government. This last feature of Roman society ts so critical a test of my

hypothesis that I shall consider it separately in a later chapter. | How did the Roman conception of slavery respond to this environment? Did the acknowledged superiority of Greek culture create an exception to the rule I am here maintaining, that slaves are always regarded as persons without honor? To answer these questions, the legal system is as good a place to begin as any; for it is here that we find the traditional notions of dignitas or honor not only persisting, but acquiring new significance by the period of the late republic, as Peter Garnsey demonstrates in his excellent study of social status and legal privilege in ancient Rome. The entire legal system, Garnsey shows, was based on the principle of privilege. There was a dual legal structure, one for those who had privilege and another for those who did not. The privileged were tried in a different court, and the penalties they received differed from those meted out to the nonprivileged who had committed the same offense. There were several channels of privilege; these included birth, Roman citizenship, wealth, and proximity to power. However, the main channel of legal privilege was “the possession of honor or dignitas, which derived from character, birth, office, and wealth.”°’ Dignitas, according to Cicero, is “honorable prestige. It merits respect, honor and reverence.”°* And, Garnsey elaborates, “emphasis is placed on moral qualities, manner of life and the esteem which they evoke—or rather command.”®”

The Greeks, we have seen, had their own highly developed sense of honor and we know that the Romans greatly respected their civilization. In Horace’s famous phrase, “Captive Greece held her captor captive.”’° Modern historians attest to the accuracy of this aphorism. Chester G. Starr, to cite a typical view, marvels that “the degree to which the conqueror bent culturally before the conquered and humbly admitted his own inferiority in thought and tongue was extraordinary.”’' Did the Roman master, then, accept the Greek slave’s highly developed conception of his

own honor?

He did not. However much the Roman master admired his Greek slave, the one thing he always denied him was a confirmation of his sense of honor. Indeed, he went further: he denied the very existence of honor in Greek culture, seeing this as one of its major failings. The opening passage of Garnsey’s work states that the Romans viewed punishment not only as a deterrent and correction, but as something aimed at “the preservation of honor, when the dignity and prestige of the injured party must be protected, lest, if the offense is allowed to go without punishment, he be brought into contempt and his honor be impaired.”’* Of special interest in this statement is

90 The Internal Relations of Slavery

the fact that the Romans made a point of contrasting their own view of punishment with that of the Greeks. Plato’s theory, they argued, however admirable in other respects, was flawed by its failure to recognize the principle of honor as the crucial element in the infliction of punishment. It is one thing, however, to comment from afar on the intellectual products of another people and quite another to hold a similar view about them face-to-face. Just what was the attitude of the Romans toward their Greek slaves? The closer the relationship, it seems, the greater was the tendency to

deny the quality of honor to the vanquished Greeks. Contact between Romans and Greeks went back to Etruscan times; still, the available data, which are quite limited, suggest that until the end of the fourth century the prevailing feeling was one of mutual indifference.’’ As late as 200 B.c. the Roman state had no eastern policy to speak of. Understandably, Greek attitudes began to change first. By 268 B.c. the reputation of the Romans as a people of good faith and integrity was well advanced among the Greeks.” Increasingly, Rome came to be regarded with awe and admiration. Perhaps the best expression of this sentiment was Melinno’s “Hymn to Rome,” written in the early part of the second century B.c.” The Roman attitude toward the Greeks changed in just the opposite direction. By the start of the second century B.c. “no Greek could help being distressed by the almost universal contempt shown, at least in public utterances, toward his nation.”’© In stark contrast to the identification of the Roman name with good faith among the Greeks, the term Graeca fides among the Romans came to mean uncreditworthiness.’’ The Greek classicist Nicholas Petrochitos has made a special study of Roman attitudes to-

ward the Greeks, and his findings fully support my argument.’* The Romans, he shows, soon developed a set of stereotypes about the Greeks, which centered on what they considered to be the six main failings of the Greek character: (1) volubitas, a tendency to prefer formal facility in speech to substance; (2) ineptia, a proclivity for inappropriate or excessive behavior,

a readiness to elaborate on subjects of which they knew nothing; (3) arrogantia and impudentia, related according to Cicero to “irresponsibility, deceitfulness and an aptitude for flattery”; (4) deceitfulness, singled out as a particularly unpleasant trait; (5) a weakness for excessive luxury and ostentation. But it was the sixth quality that the Romans most despised: Jevitas. Embracing “aspects of instability, rashness and irresponsibility,” it connoted “absence of good faith, honor and trustworthiness” and was “a prominent element in the popular conception of Greek character.””’ Cicero, in a celebrated case, tried to win support for his plea by impugning the credibility of the Greek witnesses on this basis, and Petrochitos comments that “levitas here is that lack of credibility which is the consequence of subordinating standards of honor and duty to personal and unworthy motives, and it is attributed by Cicero to the Greeks as a people.”*’ The Romans made a

| Honor and Degradation 91 point of contrasting the traditional Roman qualities of gravitas and dignitas

with the Greek Jevitas.

Finally, it was from the relationship between Roman master and Greek slave that the diminutive graeculus came, especially from the household context in which the Greek slave performed the role of tutor. The tutor may have been admired for his intellectual excellence, but the affection was always tinged with contempt. The term graeculus seems to have suggested “Greek unmanliness” and also “general worthlessness.” Petrochitos concludes: “Graeculus is thus a word of unique type, a diminutive formed from an ethnic name; it reflects the special quality of the relationship of Roman and Greek; by nature of being a diminutive it can express a variety of attributes from the mildly patronizing to the openly contemptuous.”*! Like the

American term “sambo” and the Jamaican “quashee,” graeculus could sometimes be a term of endearment without losing its undertone of contempt. Significantly, this insulting term first appeared in Cicero’s time, when the system of slavery was at or near its peak in the Roman socioeconomic order. The Romans, we know, were not a particularly chauvinistic people. In fact, with the possible exception of the ancient Persians, they were among the least chauvinistic peoples of all time.** A good deal of the carping at the Greek way of life sprang from Roman defensiveness about their own culture, which in many areas had benefited from Greek influence; thus the lack of honor attributed to the Greeks in their midst must have come from the master-slave relationship and the tendency to view all eastern slaves (many of whom ended up in the household) as Greeks. In other words, the causal

chain did not run from a stereotype about the Greeks as a people without honor to a stereotype of the eastern slave as a person without honor but rather in the opposite direction: Greeks as a group came to be regarded as persons without honor because the great majority of slaves in face-to-face

contact with Roman masters were either ethnic Greeks or hellenized

peoples.

That the Romans were fully aware of the distinction between free Greeks and slave Greeks is made clear in their efforts to find proper tutors for their children. Most Romans found it cheap and convenient to hire or buy a slave tutor. As in Greece, though, there was always concern about the effects of this means of education on the character of the Roman child. The problem was a favorite theme of Roman moralists, especially Juvenal. And, of course, it was the stuff of Roman comedy; the relationship between the adolescens and the servus callidus, the intriguing slave, was always the funniest part of these plays.*? The point is well illustrated by a humorous ex-

change between a free Greek teacher and a Roman father who was evidently caught in the dilemma of choosing between the quality of his son’s education and its inordinate expense, a dilemma that two thousand years of educational reform has yet to resolve:

92 The Internal Relations of Slavery oo _ “How much will you charge to teach my son?” the father asked Aristippus. “A thousand drachmae,” replied Aristippus, who obviously had a high opinion of his worth. “But I can buy a slave for that,” returned the father, to which the sharpwitted Aristippus rejoined: “Then you will have two slaves—your son and the one you buy.’”*

What held for the civilized and culturally admired Greek slaves held even more for other slaves. The typical slave in the Roman view was a “vocal instrument,” a nonperson to be used sexually, disciplined with the whip, and questioned in court only under torture since his word was utterly

without honor. At the same time the Roman master, even more than his Greek counterpart, took special delight in possessing a large retinue of slaves. Nothing more enhanced his sense of honor and his reputation in the eyes of his peers. And, as in all slave societies, even the poor who may have owned no slaves felt a sense of honor in the presence of slaves. In this sense the system was self-regulating; it “fed on itself,”’ as Keith Hopkins indicates: “The presence of substantial numbers of slaves in Roman society defined free citizens, even if they were poor, as superior. At the same time, free citizens’ sense of superiority probably limited their willingness to compete with slaves, to work full time as the overt dependents of other citizens. Yet rich men, by definition, needed dependents. Slavery permitted the ostentatious display of wealth in the palaces of the rich without involving the degradation of the free poor.”® The Roman master, it should be emphasized, demanded more than mere obedience from his slave. Seneca no doubt spoke for his class when he drew the distinction between ministerium, which is the performance by the slave of what he is obliged to do, and beneficium, which is what was performed

“not by command but voluntarily.”®° | What was true of ancient Rome held equally for the slaveholding socie-

ties of the Islamic world, especially those of the Arabs, all of whom had highly timocratic cultures.*’ Perhaps more than in any other part of the pre-

modern world, slavery there was not only a state of dishonor, but one in which a major function of the institution was to support the dignitas of the master. The modern ethnography on the Arabs and other Middle Eastern peoples presents innumerable instances of this. A few examples will suffice. Harry St. John Briger Philby, who traveled in Saudi Arabia during the 1930s, recalled how a slave, Shabban, was sent by his master with a present of two sheep for Philby’s dinner. “I naturally proffered the usual money gift,” he wrote, “and was somewhat taken aback at his absolute refusal to accept it. That was certainly unusual though very credible. He visited me several times during the afternoon, and proved to be a person who combined strength of character with a manner of great charm. And though only

Honor and Degradation 93 a slave and practically a full-blooded Negro, he seemed to exercise the authority of his master by proxy, as to the manner born” (emphasis added).** All masters, especially Arabs, desired slaves such as Shabban. When they did find one, they were prepared to protect him more carefully than their own son—for their honor was thoroughly invested in this human possession. Harold H. P. Dickson, who visited Kuwait during the early part of this century, observed: “They [the slaves] know that their lord will avenge himself on any stranger who harms his slave, more than if he were his own son. This is literally true, for to kill or kidnap a man’s slave affects his honor, not so the slaying of his son.”*’ And as Sylvia Bailes, who cites this passage, adds: “It should be noted, of course, that the Kuwait master is not actually

concerned about his slave, but rather about his honor.””? Even where the Arabs developed large-scale plantation slavery, as they did in Zanzibar, the psychosocial significance of slaveholding remained at least as great as its economic value. As Frederick Cooper points out, this so puzzled European observers that they branded the Arab planters as indolent and unambitious.’! Economic success, however, was only one element of the highly timocratic slave-based culture of the Arab elite. The term that was central to the Zanzibari elite’s definition of its identity was heshima, which

meant “respect.” In more concrete terms heshima meant, in addition to being prosperous, having a large retinue of slaves, a good family background, and an aristocratic demeanor. “Having dependent followers had long been an important component of power and prestige. The increasing economic importance of slaves was added to their social value—whether the clove industry prospered or stagnated, the slaves’ labor helped provide subsistence while their presence conveyed prestige.””” Cooper, in emphasizing the paternalistic ethos in Zanzibar, the integrative tendencies of the society, and the manifest ideology of Islam (which defined the slave as socially inferior but equal in the sight of God), comes uneasily close to neglecting the inherent degradation of slave status.”* Slavery was more than simply “subordination”; it was considered a degraded condition, reinforced by racist attitudes among the Arab slave owners.” In Zanzibar, as in all other Arab states, it therefore was forbidden to enslave fellow Arabs. Roger F. Morton, in his study of Arab slavery on the Kenya coast, emphasized the role of racism in the master’s view of slaves, and there is no reason to believe that the Arabs of Zanzibar were any different: “Rendered

inferior by birth, occupation, and color, slaves became natural objects of abuse for the Muslim free-born.””’ Furthermore, what Finley says of ancient Greece and Rome holds equally for Zanzibar and the Kenya coast during the nineteenth century: whatever the relationship between masters and slaves, the fact that slaves could be and were subject to corporal punishment was an implicit statement of their degradation. In Muslim East

94 The Internal Relations of Slavery Africa, as in the ancient world, slaves were (with very few exceptions) the only category of persons subject to whipping by private persons. In East Africa, as in the U.S. South, their degradation meant that all free persons regarded them as fit subjects for abuse. It was not just in Georgia that slaves, once outside the protective power of their masters, found themselves victims

of mob violence; it happened also in Malindi and Mombasa.” There can be no doubt that the slaves themselves regarded their condition as a degraded one. While they have left little verbal account of their feelings, their actions speak directly to the point. In spite of the paternalism of the Zanzibari and coastal Kenyan slave orders, slaves ran away in droves, risking severe exploitation by the strange peoples of the interior and even more murderous retaliation by their masters if and when they were recaptured.”’

Honor and Slavery in the U.S. South What is true of premodern slavery holds equally for the slave regimes of the Americas. E. D. Genovese’s claim that “the old South came closest of all

New World slaveholding regimes to producing a genuine slave society” might be open to challenge, but he certainly is correct in arguing that the master class of the Old South developed to its highest degree a slaveholder’s ideology.”®

By no accident did this ideology expand into the most elaborate and deliberately articulated timocracy of modern times. One part of the ideology referred to the master’s own conception of himself, and it is generally agreed that its pivotal value is the notion of honor, with the attendant virtues of manliness and chivalry. The historian Clement Eaton, himself a southerner, writes of the southern slave plantocracy that “despite their faults the Southern aristocracy had resplendent virtues that seem archaic in our industrial society today—their code of personal and regional honor, their devotion to a cause and their appreciation of chivalric conduct.””’ It was not in spite of, but because of their faults, especially their slavedriving, that they possessed these “resplendent virtues.” The same held for their love of freedom. When Samuel Johnson asked, “How is it that we hear the loudest yelps for liberty among the drivers of negroes?””'”’ he betrayed a rare failure of irony, not to mention a superficial grasp of the history of the idea of liberty. There was nothing at all hypocritical or anomalous about the southerner’s highly developed sense of honor and freedom. Those who most dishonor and constrain others are in the best position to appreciate what joy it is to possess what they deny. It is important to emphasize the connection between slavery and timocracy in the Old South, for the link is sometimes denied by historians. Rollin G. Osterweis argues correctly that “the civilization of the Old South rested on a tripod—

Honor and Degradation 95 cotton and the plantation system forming one leg, Negro slavery a second,” and what he calls “the chivalric cult” constituting the third.'°' He correctly identifies its essential features: an excessively developed sense of honor and pride, militarism, the idealization and seclusion of women, and regional nationalism. While he recognized the role of slavery in the development of this cult, he claimed that its significance lay mainly in providing the planter class with enough leisure to be receptive to romantic ideas from Europe. Thus he interprets the chivalric cult as “a manifestation of the romantic movement.” European ideas no doubt contributed to the southern intellectual expression

of its timocratic culture, but it is to reify ideas excessively to claim that southern timocracy was a manifestation of European romanticism. Even though it is not possible to authenticate any direct causal relationship between slavery and timocracy in the slave systems of Greco-Roman antiquity, there can be no doubt about the direct link between the two in the culture of the slave South. One of the most definitive statements of the relationship is provided by John. Hope Franklin. First, he correctly emphasizes the notion of honor—not romanticism—as the central, articulating principle of southern life and culture. It was something inviolable and precious to the ego, to be protected at every cost. It promoted extravagance, because of the imputation of poverty which might follow retrenchment. It sanctioned prompt demand for the redress of grievance, because of the imputation of guilt that might follow a less precipitate policy. It countenanced great recklessness of life, because of the imputation of cowardice that might follow forgiveness of injuries. The honor of the Southerner caused him to defend with his life the slightest suggestion of irregularity in his honesty or integrity; and he was fiercely sensitive to any imputation that might cast a shadow on the character of the women in his fam-

ily. To him nothing was more important than honor. Indeed, he placed it above wealth, art, learning, and the other “delicacies” of an urban civiliza-

tion and regarded its protection as a continuing preoccupation.” , Second, Franklin shows how the notion of honor diffused down to all free members of the society from its ruling-class origins.'”° Third, and most important, he demonstrates the direct causal link between the southern ruling class’s excessively developed sense of honor and the institution of slavery.“ More specifically, he shows how the master’s sense of honor was derived directly from the degradation of his slave, beginning in childhood and continuing through life in his despotic exercise of power. Franklin leaves no doubt concerning the veracity of Thomas Jefferson’s observation that the relationship between his fellow slave owners and their slaves was “a perpetual exercise of the most boisterous passions, the most unremitting despotism on the one part; and degrading submission on the other.”!” Nor was the connection lost on other, less celebrated southerners. The Alabama lawyer Daniel R. Hundley, for example, wrote in 1860 that “the

96 The Internal Relations of Slavery natural dignity of manner peculiar to the Southern Gentleman, is doubtless owing to his habitual use of authority from his earliest years,” and he goes on to give the classic rationalization for this means of achieving “dignity” when he adds, “for while coarser natures are ever rendered more savage and brutal by being allowed the control of others, refined natures on the contrary are invariably perfected by the same means, their sense of the responsibility and its incident obligations teaching them first to control themselves before attempting to exact obedience from the inferior natures placed under their charge.”'°° Perhaps the most brutally cogent statement of the inexorable link between southern slavery and timocracy was given by the Confederate soldier who described his flag as the symbol of “an adored trinity—cotton, niggers and chivalry.” '°’

The other side of southern timocracy was the ideology of “Sambo,” the degraded man-child that, to the southerner, constituted the image of the slave. Stanley Elkins summarizes this stereotype as follows: “Sambo,

the typical plantation slave, was docile but irresponsible, loyal but lazy, humble but chronically given to lying and stealing; his behavior was full of infantile silliness and his talk inflated with childish exaggeration. His relationship with his master was one of utter dependence and childlike attachment: it was indeed this childlike quality that was the very key to his being.

Although the merest hint of Sambo’s ‘manhood’ might fill the Southern breast with scorn, the child, ‘in his place’, could be both exasperating and lovable.”'°® As a description of how the typical southern master felt about his slave, this is quite accurate. An almost identical stereotype of the slave existed in the Caribbean.'”? And how reminiscent it is of the ancient slaveholder’s Graeculus conception of his slave. The stereotype is, in fact, an ideological imperative of all systems of slavery, from the most primitive to the most advanced. It is simply an elaboration of the notion that the slave is quintessentially a person without honor. The key to Sambo, Elkins rightly notes, is the total absence of any hint of “manhood,” which in turn is a perfect description of the dishonored condition. The existence of the Sambo ideology in the South, as in all other slave systems, is further proof of my claim that slaves are universally treated as dishonored persons. The Sambo ideology, however, is no more realistic a

description of how slaves actually thought and behaved than was the inflated conception of honor and sense of freedom an accurate description of their masters. What was real was the sense of honor held by the master, its denial to the slave, its enhancement through the degradation of the slave, and possibly the slave’s own feeling of being dishonored and degraded. Beyond this it is difficult to generalize, for the degree to which a master class is prepared to defend its honor against rebelling slaves or invading outsiders, and the degree to which a slave population will accommodate to or reject its dishonored condition, are functions of the peculiar structure, inter-

Honor and Degradation 97 nal strength, and external constraints of the slave system in which they find themselves. In some cases, as among the slaveholders of the emirates of northern Nigeria during the nineteenth century, the sense of honor and its peculiarly Islamic elaboration proved to be highly functional in perpetuating the system.''® In other instances, such as the antebellum South, the exaggerated sense of honor and quixotic chivalry of the ruling class proved to be the major cause of its undoing. Among some slave populations the condition of dishonor as well as the sense of being so dishonored might, given the right revolutionary opening, prove to be an important asset in the struggle for emancipation, as it was in parts of the Greco-Roman world during the second and first centuries B.c.,''' in the dead lands of lower Mesopotamia during the late ninth century,''” and in many areas of the Caribbean and Latin America between the seventeenth and nineteenth centuries. In countless other situations the viability of the slave system, the solidarity of the ruling class, and the absence of revolutionary openings dictate that the slave population nurse its sense of dishonor, accept its dishonored condition, or find alternative means of expressing or sublimating its grievances.'!” There is absolutely no evidence from the long and dismal annals of slavery to suggest that any group of slaves ever internalized the conception of degradation held by their masters. To be dishonored—and to sense, however acutely, such dishonor—is not to lose the quintessential human urge to participate and to want a place.'!* Indeed, it is precisely this irrepressible yearning for dignity and recognition that is hardest to understand about the condition of slavery. The fundamental problem posed by slavery may be simply stated as one of incentive and mutual recognition. The master not only forces the slave to serve him with the threat and the actuality of physical violence, he heaps insult upon injury by continually degrading him. Why does the slave obey? Why does the master so wantonly appear to undermine his own best interest by degrading his slave? What really is going on?

| Hegel and the Dialectics of Slavery It is this fundamental dilemma that so intrigued Hegel.''? An examination of his analysis is instructive not only for the profound insights it offers, but for what we can learn from a critique of its limitations. The master’s domination of the slave is seen by Hegel as a paradigm of inequality: “The one is independent whose essential nature is to be for itself, the other is dependent whose essence is life or existence for another. The former is Master, or Lord, the latter is Bondsman.”''® The master’s existence is enhanced by the slave’s, for in addition to existing on his own account his consciousness is mediated through another consciousness, that of the slave. In other words, another person lives through and by him—becomes his sur-

rogate—and the master’s power and honor is thereby enhanced. The mas-

98 The Internal Relations of Slavery ter’s independence becomes the real—the only—basis of the slave’s thralldom. By negating the slave’s existence, the master seems to solve one of the most pressing problems of a free and equal relationship: the frustration that the other, if he is free, is also strongly desirous of winning confirmation of his identity from ego. Both are struggling to gain the other’s confirmation of their superior identity. All free relationships amount to a “life-and-death struggle.” Slavery appears to solve this dilemma. The slave cannot negate the master, for whatever he does is done on behalf of his master. The slave dies, it is true, but he dies in the master; so the master becomes autoconfirming, so to speak. But this one-sided and unequal form of recognition soon reveals its limitations. At precisely the point where the master achieves lordship, he finds that he has become dependent on his slave. He cannot be sure even of his own existence, since the reality of his domination rests on the unreality of that which he masters: the slave, whom he has socially killed and rendered non-essential by making him merely an extension of himself. Further, the slave cannot confirm his honor, cannot offer recognition, because he is not worthy. This is what Alexandre Kojéve, in his celebrated commentary, calls the master’s “existential impasse.”''’

The opposite, Hegel thought, was true of the slave: “Just as lordship showed its essential nature to be the reverse of what it wants to be, so too, bondage will, when complete, pass in the opposite of what it immediately is.”''® The slave, by his social death, and by living “in mortal terror of his sovereign master” becomes acutely conscious of both life and freedom. The idea of freedom is born, not in the consciousness of the master, but in the reality of the slave’s condition. Freedom can mean nothing positive to the master; only control is meaningful. For the slave, freedom begins with the consciousness that real life comes with the negation of his social death. (What I am here calling the negation of social death is what Hegel, with his usual verbal extravagance, calls “extraneous alien negation.) Freedom— life—is a double negation; for his condition is already a negation of life, and the reclamation of that life must therefore be the negation of this negation. Nevertheless, freedom is more than just a double negation. It is continuously active and creative. The slave, in his social death, is already once transformed. The life he strives to regain cannot be the life he lost. In his enslavement the slave has become a new man for his master; in his struggle for freedom and in his ultimate disenslavement the slave, Hegel believes, becomes a new man for himself. And here is the most surprisingly radical insight in all of Hegel’s work, the insight that was to have so profound an effect on Marx and on generations of subsequent radical thinkers.''? How does the slave become positively free? How does he make a new man of himself? “Through work and labour,” answers Hegel, “this consciousness of the bondsman comes to itself’; for labor “‘is desire restrained and checked,

Honor and Degradation 99 evanescence delayed and postponed; in other words labour shapes and fashions the thing.’ Consciousness, through work, creates object, becomes externalized, and passes “into something that is permanent and remains. The consciousness that toils and serves accordingly comes by this means to view that independent being as its self.” He adds, by way of conclusion, “Thus precisely in labour where there seemed to be merely some outsider’s mind and ideas involved, the bondsman becomes aware, through this rediscovery

of himself by himself, of having a being and a mind of his own.”'”° | Hegel is partly night and partly wrong in arriving at this conclusion. Ironically, he is wrong precisely where most commentators, including Marx and Kojéve, have considered him most insightful. There is nothing in the

nature of slavery which requires that the slave be a worker. Worker qua worker has no intrinsic relation to slave qua slave. This does not mean that the slave cannot be used as a worker. Indeed, his slaveness, especially his natal alienation, made possible his effective exploitation as laborer in conditions where no other kind of laborer would do. But this does not in any way mean that slave necessarily implies worker. I have repeatedly stressed that most slaves in most precapitalist societies were not enslaved in order to be

their masters.

made over into workers; they may even have been economic burdens on Further, I disagree totally with the view that slavery created an existential impasse for the master. In the first place, the master could and usually did achieve the recognition he needed from other free persons, including other masters. In almost all large-scale slaveholding societies, not to mention those in which slavery was not structurally important, there was a sizable class of free nonslaveholding persons; indeed they usually constituted the majority of all persons in such societies. As we have seen, the nonslaveholding free group invariably came to adopt elements of the timocratic character of the master class. The poorest free person took pride in the fact that he was not a slave. By sharing in the collective honor of the master class, all free persons legitimized the principle of honor and thereby recognized the members of the master class as those most adorned with honor and glory. Beyond this, the degradation of the slave nurtured the master’s sense of honor, both in his childhood training by slave nannies and, throughout his life, as a ready object for the exercise of his sense of power.'*! In a small but important minority of large-scale slave societies, however, almost all free persons were masters. This was true of the total slave systems of the Caribbean and the equally brutal though isolated instance of Dutch East Indian slavery in the Banda group of the Spice Islands to the south of Ceram.'* In these societies we do find something approaching Hegel’s crisis of honor and recognition among the master class. Other than themselves, there was only the brutally used and utterly despised class of slaves to recognize the prestige of the masters. Faced with this dilemma, the master class

100 The Internal Relations of Slavery did two things. In the slave society where they procured their wealth, they abandoned all claims to honor and any attempt to develop a timocratic culture. Slavery, they recognized, degraded both master and slave. Hence they dropped all pretensions to culture and civilization and simply indulged their appetites. The slave women whom they whipped in the fields during the day became their bedmates during the night. There were no attempts to build great manors decorated with idolized wives. More often than not, the mistress of the large stone hovel that passed for a great house in the Caribbean was herself a slave. Since there was no one to confirm honor, it was simply thrown to the winds. But there was another solution. The successful master, as soon as he made his fortune, would pack up and flee the degraded source of his wealth.

He would return to Europe, where he could ostentatiously display his wealth, proclaim his honor, and have it confirmed by the free population of the metropolis. It is this which largely explains the high rate of absenteeism among successful members of the planter class in the Caribbean.!” Thus in criticizing Hegel’s failure to take account of the free nonslaveholding members of the master’s society, we arrive at an extremely important, if paradoxical, conclusion about the nature of slave-based timocratic cultures: namely, that they are possible only where slavery does not totally dominate the society. A truly vibrant slave culture, if it is to avoid the crisis of honor and recognition, must have a substantial free population. Conversely, a society with only masters and slaves cannot sustain a slave culture. Leaving aside such extreme cases as the British Caribbean slave societies and those of the Banda Islands, we must still answer the very basic questions set forth earlier about the reasons for the slave’s degradation and the benefits to the master. I have said that Hegel was partially right in the answer he gave; specifically, in that he pointed to one solution to the dilemma. Confronted with the master’s outrageous effort to deny him all dignity, the slave even more than the master came to know and to desire passionately this very attribute. For dignity, like love, is one of those human qualities that are most intensely felt and understood when they are absent—or unrequited. Slavery, for the slave, was truly a “trial by death,” as Hegel called it. Out of this trial the slave emerged, if he survived at all, as a person afire with the knowledge of and the need for dignity and honor. We now understand how very superficial are assertions that the slave internalized the degraded conception of him held by the master; or that his person was necessarily degraded by his degraded condition. Quite the opposite was the case, Hegel speculated, and what evidence there is fully supports him. Thus whenever we hear the voice of the slave himself, or whenever we hear from chroniclers and analysts who attempt to probe behind planterclass ideology into the actual feelings of the slave, what invariably surfaces

is the incredible dignity of the slave.’ |

Honor and Degradation 101 This leads us to one of the most remarkable features of slavery. What does the master make of the slave’s yearning for dignity, itself part of his wider yearning for disalienation and relief from the master’s all-embracing power? In all but a handful of slaveholding societies the master exploits this very yearning for his own benefit. How? He does so by manipulating it as the principal means of motivating the slave, who desires nothing more passionately than dignity, belonging, and release. By holding out the promise of redemption, the master provides himself with a motivating force more powerful than any whip. Slavery in this way was a self-correcting institution: what it denied the slave it utilized as the major means of motivating him. The slave constantly struggled with his master for recognition, for survival. Somewhat like Saint Augustine, who found in the very depth of grief over his friend’s death that he was “at once utterly weary of life and in great fear of death”'*° and in this way came to love life even more, so did the slave

in the weariness of his degradation and social death come to a passionate zeal for dignity and freedom. The dialectic does not end here. The slave’s struggle made it necessary that the master, in order to make slavery workable, provide an opportunity for the negation of slavery. The conflict between master and slave became transformed from a personal into an institutional dialectic, in which slavery, as an enduring social process, stood opposite to and required manumission as an essential precondition. How did this come about? What were the institutional mechanisms that brought slavery into being and, in order to sustain it, generated further mechanisms for its negation? And what of that handful of societies that resisted this institutional dialectic and denied manumission? Why did they reject this resolution? How was the dignity of the slave, which sprang from his degradation, expressed and contained in such systems? To an exploration of these and related problems we now turn our attention.

TT

SLAVERY

AS AN

INSTITUTIONAL PROCESS

_ Enslavement

4 Ih

of “Free Persons

IN THE STUDY of the sources of slaves two closely related but separate issues are invariably confused: the problem of how persons be-

came slaves, and the problem of how slaveholders acquired slaves. The means by which persons were enslaved are legion and include many that were peculiar to certain societies. The overwhelming majority, however, may be grouped under eight heads: (1) Capture in warfare (2) Kidnapping (3) Tribute and tax payment (4) Debt (5) Punishment for crimes (6) Abandonment and sale of children (7) Self-enslavement (8) Birth.

The first seven means, involving persons who were born free and subsequently were reduced to slavery, will be considered in this chapter. Birth, the

most important method of enslavement, will be discussed separately in Chapter 5. Then, in Chapter 6, we shall look at the various means of acquisition. 105

106 =‘ Slavery as an Institutional Process

Capture in Warfare Throughout history, captivity in warfare has been one of the major means by which persons have been reduced to slavery. It is easy, however, to exaggerate the role of warfare as a source of slaves. If we are to place captivity in warfare in proper perspective, we need to clarify certain important and neglected issues. The first is the distinction between captivity in warfare as a current means of enslavement and as an original means of enslavement. By

current means of enslavement, I mean its relative significance for a slave population at a given moment in time. By original means of enslavement, | refer to its role in the enslavement of the ancestors of a slave population observed at any given moment. The complicating factor is enslavement by birth. Captivity in warfare, even when a major factor in the enslavement of

creased. ,

the ancestors of all persons born as slaves, usually declined in relative significance as the proportion of the slave population enslaved through birth in-

A second important point is that captivity in warfare should not be confused with enslavement by means of such captivity. This is a major problem in the literature on slavery, largely because of the frequent and wholly erroneous assumption that the fate of most prisoners of war was enslavement. It is simply not true that the majority of persons captured in warfare have been enslaved, even if only premodern societies are considered. What is more, it 1s incorrect to assume that even if a society keeps slaves in great numbers it will enslave all or most of its captives. There are several reasons for this. One is the logistics of warfare. Having a large number of prisoners is an encumbrance for an army in the field. Even if it was decided to profit from the enslavement of prisoners, it could still be a formidable problem for soldiers to return home with a batch of chained slaves. The best course of action was to sell the prisoners to traders as soon as possible, even when there was a strong demand for slaves in the home society of the victors. Accounts of ancient warfare and slavery sometimes foster the misleading impression of Roman, Carthaginian, and Grecian officers marching home accompanied by thousands of slaves, with eager expectations of employing them on their latifundia or home farms. This must rarely have happened, even where the officers owned slaves. Before assessing the role of warfare as a means of enslavement, let us consider the more common experiences of captives. The alternatives were immediate massacre; torture and sacrifice, sometimes culminating in cannibalism; ransom; prisoner exchange; temporary imprisonment; serfdom; impressment in the victor’s army; colonization; and simple release. There is no relation between the level of development of a victorious

group and its treatment of prisoners. The Tupinamba and the Aztecs, for instance, differed vastly in sociopolitical complexity but treated their prison-

ers in much the same way, both engaging in highly ritualized, sadistic

Enslavement of “Free” Persons 107 slaughter and cannibalism.'! Torture, massacre, ransom, and all the rest might be employed along with enslavement. This was true not only of primitive tribes such as the precontact Cherokees and other Indians of southeastern North America, as well as those of the northwestern coast,” but of highly advanced peoples. The Carthaginians ritually sacrificed thousands of prisoners.” The ancient Greeks and Romans throughout their history massacred not only soldiers in the field but defenseless inhabitants of captured cities. Human sacrifice of prisoners was practiced occasionally by the Romans from 225 B.c. until its abolition by the Senate in 97 B.c., and even was recognized as a form of ritual sacrifice.*

Ransom, however, was the more common fate of prisoners of war. Among the Nkundu and the Luvale of central Africa, the person who was primarily responsible for starting a war had to find the means of ransoming anyone taken during hostilities, otherwise he himself ended up as the slave of the aggrieved kinsmen, a salutary check on adventurous spirits.” Among all the advanced states of Africa, Asia, and Europe upper-class captives were usually ransomed. As one would expect, the higher the rank of the captive, the greater was his ransom. Sometimes this could be quite excessive: during the Third Dynasty of Ur vast and ruinous sums were asked for upper-class officers.° Among many peoples prisoners were taken mainly with the intent of exchanging them for their own members taken in previous battles, or for purely commercial purposes: Icelandic warriors made a tidy sum by this means,’ and the Kerebe of Tanzania took slaves mainly to exchange them for cattle, which interested them far more than the “two-footed stock” preferred by the Greeks.* Warfare for the Margi of northern Nigeria “served largely as a wife-recruiting and ransom collecting institution.”” Among Islamic peoples captured coreligionists were, by law, offered for ransom and not enslaved, but this regulation was not always observed and even if it were, negotiations sometimes went on interminably.'° Between Christians and Muslims, of course, religious scruples worked against the captive.'! From medieval through early modern times the North African Muslim states relied on captives as an important source of both income and slaves. In early modern Algiers, according to Ellen Friedman, “the labor services of captives as well as ransoms paid for them were critical to the Algerian econ-

omy.”

In many advanced premodern societies prisoners of war were incorporated into the victors’ societies in a dependent status other than slavery. This tendency was most marked in the ancient Near East and Orient. The practice existed side by side with the institution of slavery. Indeed, for this very reason it was long assumed that in ancient Mesopotamia and China both, all prisoners of war were automatically enslaved. Many authorities still make this assumption,'’ but it has been strongly challenged by recent scholarship. Because China and Mesopotamia are the two best-known and most controversial cases, it is worth examining them at some length. Chinese scholars are divided on the destiny and role of the vast numbers

108 Slavery as an Institutional Process of prisoners taken in the many wars of the ancient Chinese. What may be called the “hard-line” periodization view, best represented by the works of Chou ku-Cheng,"* holds that large-scale slavery existed in ancient China and that the vast majority of slaves originated as prisoners taken from defeated clans. On the whole, however, even those Marxist scholars who insist on a period of large-scale slavery tend to argue against the view that prisoners of war were the major source of slaves. Kuo Mo-jo, the most eminent of the Marxists, is rather vague on this issue in his frequently cited discussion of the role of sacrifice.'” Less equivocal is Tung Shu-yeh who, while claim-

ing that slaves at one period constituted as high as 25 percent of the total population, nonetheless asserts that very few were recruited from among the prisoners of war—crime and debt being the main sources.'® Finally, among Communist scholars, there is the lengthy discussion by Chien Po-tsan who, against his colleagues, promulgates the view that there could not have been large-scale slavery during the eastern and western Han dynasties because prisoners of war were not available in quantity. Wars with the barbarians, he argues, were fought to establish commercial routes through central Asia, to expand the empire, and to force the conquered to pay tribute. The taking of prisoners of war was incidental, engaged in principally as a form of intimidation and revenge. It was highly exceptional for prisoners to be given to officers as a form of payment or encouragement.'’ Western scholars tend to disagree with almost all these interpretations. The standard position is that of C. M. Wilbur.'® He observes that while in

the former Han dynasty “thousands of the enemy were captured ... it cannot be lightly assumed that these prisoners were enslaved.”'’ He concludes: What became of the thousands and probably hundreds of thousands of prisoners of war is a historical enigma. Some were enslaved but there is no evidence of this on a large scale. Han histories simply neglect to tell what hap-

pened to the prisoners taken in wars against the Hsuiung-nu, against the oasis states of the northwest, the Koreans, and the kingdoms of south China. It is not even possible in most instances to distinguish numerically or proportionately between enemy slain and captured. This seems significant: it was apparently a matter of indifference to the state whether enemy soldiers were captured or killed. This would hardly have been the case if prisoners of war had been economically important as a source of slaves.”°

FE. G. Pulleyblank throws some light on the enigma by emphasizing the fact that “no sharp distinction’ was made between prisoners of war and ordinary convicts, especially during the Han and earlier periods. He notes, further, that prisoners of war were not sold by the state but were given away to officials.*' He is less inclined than Wilbur to think that they were not enslaved in substantial numbers. The most likely conclusion is that the Chinese treated prisoners of war differently during different periods of their vast history. Up to the end of the

Enslavement of “Free” Persons 109 Han dynasty the most likely practice was to enslave only a minority and to use the rest as colonists or for other purposes. After this, there was an increasing tendency to enslave the great majority of prisoners of war. By the period of the Northern dynasties (A.D. 386-618) “significant evidence is available to leave little doubt that ... enslavement of captured or surren-

dered enemies, on combat duties or otherwise, was common practice.” Prisoners of war were still used for other purposes: they were “placed in bonds, regrouped and then impressed into the victorious army, [and] settled in the victor’s sparsely populated areas,” but the evidence is clear that “‘conquered civilians were reduced to slavery en masse.””° An almost identical problem is faced by students of early Mesopotamia. Isaac Mendelsohn argues the traditional view that the vast majority of prisoners of war were enslaved, although he notes also that there were alternate uses.** Until a decade or so ago Russian students of Mesopotamia, under the dominant influence of academician V. V. Struve, dogmatically assumed that all prisoners of war were enslaved and that this was the primary source of slaves in the presumed large-scale slave systems of the area.’? In recent years, however, most Russian scholars have done a complete about-face on the issue. I. M. Diakonoff holds that most male prisoners of war in early Sumer times were killed and that they were never employed as slaves in significant numbers during later periods.”° I. I. Semenov is typical of the most extreme reaction against Struve and the Stalinist school. ““What happened to them?” he asks of the prisoners of war in ancient Mesopotamia, then anSWEIS:

In our view, a direct, unconditional identification of prisoners of war with slaves is erroneous. Prisoners of war, in and of themselves, are not yet slaves: they are still only people torn out of the system of relationships existing in

the society to which they belonged, and thus separated from the means of production ... In the ancient Eastern societies, prisoners of war, when there were more of them than could be used in domestic and auxiliary work, were usually settled on the land.”’

Most Western scholars adopt much the same position. I. J. Gelb argues forcefully that it was simply not practical to enslave prisoners of war in ancient Mesopotamia. Instead, after being branded and kept for a short period as prisoners, “they were generally freed and resettled or utilized for specialized purposes of the crown, such as the personal guard of the king, mercenaries, and a movable force.””® Gelb, however, may have gone too far in the opposite direction in denying altogether the use of prisoners of war as slaves on a significant scale.” The true situation would seem to be more like that which existed over time in China. At all times some prisoners of war were used as slaves, but from the earliest period on which we have records up to the end of the Baby-

110 ~~ Slavery as an Institutional Process

lonian dynasties only a small proportion were reduced to slavery. However, as in China, an increasing number were enslaved with the passing centuries, and by the neo-Babylonian period there is reason to believe that the majority were being enslaved. The Soviet scholar Dandamayev may have overstated his case somewhat in claiming that large-scale slavery existed in the Achaemenid empire during the sixth century B.c., although the evidence he presents strongly suggests that at that time slave labor was heavily used on the domains of the Persian upper class throughout the empire.”° Neither China nor the ancient Near Eastern states ever developed largescale slave systems, so an appropriate question is whether the fate of prisoners of war was substantially different where the combatants came from societies that relied heavily on slavery. Pierre Ducrey’s study of the treatment of prisoners of war among the ancient Greeks is instructive in this regard.*! Ducrey examined 120 cases of warfare involving Greeks from earliest times to the Roman conquest. He found 24 cases of massacre, 28 cases of general enslavement, and 68 cases in which prisoners seem to have suffered nothing more severe than simple detention followed by release.*” One must be careful in generalizing from these figures; they tell us nothing about the number of persons involved in each engagement, and the limitations of the data and the sample are obvious. Even so, it is surprising that in less than a quarter of the cases mentioned in the sources are captured soldiers sold into slavery. In spite of the tremendous demand for slaves in the ancient Greek world, the pattern of treatment of prisoners seems not to have departed significantly from what prevailed in imperial China and the ancient Near East. Ducrey draws attention to an important distinction in the discussion of prisoners of war: between soldiers captured in open engagement, and the de-

fenders and citizens of a captured city.’ The ancient sources, if not read carefully, give the impression that the inhabitants of conquered cities were routinely reduced to slavery or carried off en masse to slave markets. Ducrey found, however, that many of these claims were either improbable or too vague for us to be sure just what happened to the inhabitants of the captured cities. Where the sources are clear, the general conclusion to be drawn is that when a Greek city was besieged, its survival as a political entity was at stake rather than the survival or liberty of its citizens. In most cases the attackers and the besieged arrived at an accord that did not involve enslavement. The Greek experience points to an important general tendency in the history of enslavement: there is a strong tendency on the part of a conquering group not to enslave a conquered population en masse and in situ. This, however, is no more than a strong tendency, to which there have been many exceptions. The exceptions bring us to a second generalization, which can be stated in much stronger terms: attempts by a conquering group to enslave a conquered population en masse and in situ were almost always disastrous failures.

Enslavement of “Free” Persons 111 When a people was conquered, it was by definition the conquerors who were the outsiders to the local community and the conquered who were the natives. In this situation one of the fundamental elements of slavery—natal alienation—was almost impossible to achieve either intrusively or extrusively. By the nature of the case, the conquered native population could not be natally alienated in intrusive terms, for it was the master class who would be the intruders. It was equally difficult to natally alienate a native population extrusively, since in this case the moral community (insofar as it existed) was defined by the conquered. A community could hardly be expected to accept the idea of itself as having fallen. Nor was the strategy of divide and partially enslave likely to work, because such an act was likely only to make heroes of those selected for enslavement. There were other, purely practical reasons why the attempt to enslave a native population was likely to fail. First, there was the solidarity of the bulk of the population with the enslaved. Second, the enslaved were on their own social and physical ground and could easily survive by running away. Furthermore, they were more likely to find refuge in this situation than in the more usual cases where they were genuine outsiders or defined as morally fallen. Most conquering groups were sufficiently aware of these problems not to attempt to enslave a conquered group in situ. If a conquering elite wished to maintain (or introduce) on a large scale the institution of slavery, there was a

variety of options. One was to take over the slave population of the conquered group, if such a slave group already existed. A second option was to

bring in slaves from the outside and deliberately refrain from locally enslaving the conquered population. The most dramatic instance of this was perhaps the Dutch policy toward the conquered Khoikhoi of South Africa during the period of large-scale slavery from the late seventeenth century to

the early nineteenth.” ,

The Greeks employed all these measures and more. Perhaps the most typical strategy was the behavior of the Spartan Kallikratidas after he had captured the Athenian garrison at Methymna on Lesbos. He sold all the captured soldiers as well as the captive slaves into slavery, but freed the citi-

zens.°” In more extreme cases the entire population would be either deported or sold away into slavery, and new colonists brought in with their own slaves. This, for example, was the fate of the Poteidaians after they surrendered to the Athenians in 430 B.c. There were many cases of large-scale

transportation of conquered populations in Sicily, as in 483 B.c. when Gelon, the tyrant of Syracuse, after destroying Megara sold the common folk into slavery but declared members of the upper class citizens of Syra-

cuse°

A similar range of strategies was employed by the Romans. The fate of the population of a conquered city depended on whether or not the city had been in revolt and on whether it was taken by storm, had surrendered before

112 Slavery as an Institutional Process being stormed, or had surrendered after the battering rams had touched its walls.*’

For our purposes the most important conclusion to be drawn from the numerous references to the conquest of cities and other types of states in Greco-Roman antiquity is a negative one. Although there are frequent references to enslavement, there is not a single unambiguous case of the mass of the free members of a conquered people being successfully enslaved in situ. The closest the ancient world came to such a situation was Spartan helotry. This is a controversial subject, and at least one reputable historian of ancient Greece has declared categorically in a widely used text that “they [the helots] were very thoroughly slaves, and are often called by the standard

Greek word douloi.” Unfortunately, Antony Andrews, who made this assertion, did not tell us what he meant by the term “slaves.”** From our definition of slavery, however, it is clear that the helots were not slaves—whatever the merits of Critias’ remark, cited as supporting evidence by Andrews, “that in Sparta the free were more free and the slaves more fully slaves than

elsewhere.”*” The distinctive feature of slavery is not the degree of oppression involved; were this the case, the British proletariat at the middle of the nineteenth century would have been as much slaves as the blacks of the U.S. South, not to mention the countless millions of the Asian rural poor. But as Finley has pointed out, the helots remained nonslaves in the collective nature of their bondage and in the fact that they were a “subject community” (emphasis added),*° which in essence means that they belonged, and had rights of birth, however attenuated, including custodial claims in their parents and children. Their status as Greeks was never lost; it was only politically suspended. Nothing better attests to their nonslave status than a com-

parison of their fate after emancipation in 371 B.c. with that of black Americans after emancipation in 1865. ““The Messenians,” Finley tells us, “were at once accepted by the Greeks generally as a proper Greek community”;*' almost a century and a half after legal emancipation, black Americans are still struggling for acceptance in the community of which they are, as a group, among the oldest members. The reluctance to try to enslave a conquered native population, I have said, is only a tendency. There have been several noteworthy attempts at mass enslavement in the annals of slavery, all ending in failure. The most sustained and, not surprisingly, the most frightening, was the European attempt to enslave the Indian populations of the Americas. In both North and South America all such attempts ended unsuccessfully, although they lasted much longer than is generally acknowledged. The decimation of the Indian populations throughout the Americas following attempts to enslave them or force them into encomienda relations and reservations is well known; it is the extent of the genocide that has been fully appreciated only in recent years.”

Enslavement of “Free” Persons = 113

Nowhere were the attempts more extreme and the consequences more disastrous than in that familiar theater of European imperial horrors, the Caribbean. Hispaniola was Spain’s first colony, with a large Indian population of over a million souls when it was discovered by Columbus. In sixteen years it was reduced to about fifty thousand, by 1520 there were hardly ten thousand persons, and by 1550 under two hundred fifty.” Jamaica was even less fortunate. Its Arawak population was wiped out within a decade. The story was much the same in the other islands. Slavery, of course, was not the only cause of this destruction; disease and famine were the prime factors. Yet it is important not to underestimate the role of slavery both directly and indirectly in the demoralization and social destruction of the native population. As Kenneth R. Andrews observes: “Altogether these disruptive aspects of the conquest and exploitation of the Indians must be considered not merely as subordinate allies of the microbes but as major forces in the work of destruction.” ** This was most evident in the Central American region of the Caribbean. In Panama the notorious portage of the isthmus called the trajin resulted in untold suffering and death for the Indian slaves who carried silver and other goods for their masters.” In what kinds of slaveholding societies, then, did prisoners of war constitute the major means of enslavement? First, we find that captivity in warfare was always the most important means of enslavement among kin-based or tribal societies. This held true whether or not such societies themselves engaged in warfare, since external trading of slaves constituted one of the earliest forms of trade. Captivity in warfare remained important even where slavery became structurally very important in such societies, because descendants of slaves tended to be assimilated to nonslave status. To be sure, there were a few cases of kin-based societies with highly developed slave systems that did not rely heavily on war captives for their slaves (for example, the Toradja of the central Celebes**), but these were highly exceptional cases. More typical were the Maoris of New Zealand,*’ nearly all the preEuropean African societies with advanced slave sectors,** and the slavehold-

ing Indians of the American northwest coast.” , A second group in which captivity in warfare was the dominant means of enslavement comprises a subset of advanced societies with large-scale slavery during only the formative period of their developing slave sectors. There are two unambiguous cases in the premodern world: the ancient Greek slave systems, especially Athens, between the sixth and the end of the fifth century B.c.,” and Rome during the third and second centuries B.c.”! The pattern of change in both civilizations was complex. In Rome during the first two centuries of the empire a small though still significant proportion of the slave population would have been reduced to their condition

as a result of captivity in warfare, but this proportion may well have increased periodically from the end of the second century of our era.** In an-

114 Slavery as an Institutional Process cient Greece, too, the secular downward trend in the proportion of the slave

population reduced to their status through warfare would have been disturbed periodically during unusually unstable military situations: for example, during the Peloponnesian wars (431-404 B.c.), the unsettled period dur-

ing and after the Social War in the middle decades of the fourth century B.c., and during the early decades as well as the last half-century or so of the Hellenistic period.”” All the slave societies of the Americas, and possibly those of the capital-

istic slave systems of Mauritius and the other Mascarene Islands, fall into this group.”* They vary, of course, in the length of time during which prisoners of war dominated the slave populations. Assuming that most externally acquired slaves in the Americas before the end of the seventeenth century were captives of warfare, whether waged deliberately for slaves or not, we find that the Spanish colonies were the first to witness significant reductions in the proportion of their populations enslaved by this means, although the change to birth as the major means of enslavement was to take over two centuries.’ The colonies of North America surpassed all others in the rapidity with which enslavement by birth overtook captivity in warfare.”° Colonies of the non-Latin Caribbean showed the slowest rate of change, a factor that was to have disastrous consequences for the slaveholder class.” It cannot, of course, be assumed that even during this early period all (or even most) African slaves in the New World were taken as prisoners of war. Some undoubtedly were enslaved as a punishment for crimes, but these must have constituted a tiny proportion of exported slaves.”*® It has also been claimed that the majority of persons sold to the European slavers on the West African coast were already slaves when captured.” This argument is morally specious and in all probability factually inaccurate. Even if it were true, it merely takes the problem of the original means of enslavement one step backward, for we are still entitled to ask how these slaves were originally enslaved. It is highly likely that prior to the start of the eighteenth century the answer was captivity in warfare.” The third group of societies are those advanced systems with significant levels of slavery in which captivity was the dominant original means of enslavement and remained the dominant current means throughout the period of slavery. This was true of the Iberian Peninsula right down to the ending of slavery in early modern times.°’ Included in this group are France during Merovingian and Carolingian times,” as well as the large-scale slave systems of the Italian colonies of the Mediterranean during the late Middle Ages and early modern times (especially Cyprus, Crete, and Sicily).°* The bulk of the societies falling into this category are the Islamic slave systems, especially those of Saharan and sub-Saharan Africa, of North Africa and Muslim Spain.” The strong reliance on prisoners of war as a means of enslavement in these societies is explained by a combination of factors: the Is-

Enslavement of “Free” Persons = 115 lamic emphasis on the jihad and enslavement as a means of recruiting manpower, and the high rate of manumission requiring a constant inflow of out-

siders to replace and to increase the slave populations.” , Kidnapping Among premodern peoples it is frequently difficult to distinguish so-called

warfare from kidnapping raids by small “war parties” on neighboring groups. We have separated kidnapping from warfare on the grounds that it was not usually a communal affair and might be directed either at a neighborhood group with whom there was no overt state of warfare or at members of the kidnappers’ own group. Kidnapping was also conducted with the sole aim of acquiring captives, whereas this was often only a by-product of warfare. The distinction, however, cannot be too rigidly held. What Henry Ormerod observes of the ancient world held equally for most premodern societies: “It is... difficult to apply the modern conception of the ‘politically organized society’ to early conditions of ancient life. It was only as a result of a long process of development that the ancient world came to distinguish between foreigner and enemy, piracy and privateering, lawful trade and kidnapping.”°° In kin-based, small-scale societies kidnapping ranked a close second to captivity in warfare as a major means of enslavement, both original and current. According to I. J. Gelb, piracy and abduction (what he calls piracy slavery) represented “the main source of servile labor of ancient Mesopotamia and the Ancient Near East in general.”°’ This may be too sweeping a generalization. Gelb does not take sufficient account of persons enslaved by birth. It 1s best to qualify the statement by saying that kidnapping was the most important original means of enslavement, and continued as one of the major current means, during all periods of Mesopotamian antiquity. _ Kidnapping, especially piracy, also ranked close to captivity in warfare as an original and current means of enslavement among all the ancient and medieval slaveholding societies of the Mediterranean. The area was ideally suited to this form of enslavement, as Ormerod and others have shown.” Indeed, before modern times the area was plagued by piracy in all but one period of history, the first two centuries A.D. Individuals at sea and inhabitants of coastal towns were ceaselessly ravaged by pirates, Caesar being perhaps the most celebrated victim. The Greeks were both captors and captives in this nefarious traffic; for they captured fellow Greeks as well as barbarians. Kidnapping at sea and on land was rampant during the Persian wars. The practice subsided with the naval supremacy of Athens, which effectively policed the eastern Mediterranean during the middle decades of the fifth century B.c. With the outbreak of the Peloponnesian wars, however, kidnapping soared to new heights of atrocity. Not only were the citizens of

116 ~—- Slavery as an Institutional Process

all the belligerent states at risk, but the rights of neutrals were notoriously neglected. The heavy reliance on mercenaries simply worsened the situation, for these soldiers saw warfare, kidnapping, and privateering as equally attractive forms of employment. The situation was so bad that even Athenian generals engaged in an early form of the protection racket, guaranteeing the safety of coastal cities against kidnapping for a heavy price. Although Alexander attempted to clear the sea of pirates after 331 B.c., he was only partially successful and the effort collapsed after his death. During the last decades of the fourth century and throughout the next two centuries kidnapping became chronic, with certain peoples such as the Cretans and the Illyrians surpassing even their own notorious reputations as pirates and robbers. After the Punic wars Rome assumed the role of policeman of the Mediterranean, but until the start of the Christian era her record was marred by inconsistencies and duplicity. Indeed, perhaps the worst period in the annals of piracy and land-based kidnappings occurred during the second half of the second century B.c. and the last hundred years of the republic. Ormerod suggests that Rome’s negligence in controlling the pirates, especially those of Cilicia, was deliberate, motivated by the growing demand for slaves 1n the still expanding latifundia.”’ Only when trade came almost to a standstill did Rome act. The result was that the Mediterranean was secure from pirates and other kidnappers during the first two centuries A.D. Not so the outer seas. Arab pirates were rampant in the Red Sea throughout antiquity and medieval times, and the Black Sea coast remained infested. Even during these first two centuries, then, a substantial proportion of newly arrived slaves in Rome and other Mediterranean slaveholding societies were the victims of kidnappers.” Throughout medieval Europe kidnapping remained a major source of slaves, sometimes rivalling warfare in importance. The Vikings plagued the coastal cities of the North Sea, capturing people from one area and selling them to another, with the Irish, Welsh, and northeast Britons and Slavs being particularly subject to their raids. The massive rise in the slave population of Christian Spain during the late thirteenth, fourteenth, and early fifteenth centuries was largely made possible by piracy and privateering, which became the most important original and current source of slaves during this period.” Many of the slaves recruited to work on the large-scale sugar plantations of the Mediterranean islands from the thirteenth century must have been kidnapped, although it is difficult to distinguish them from genuine prisoners of war. They came from Greece, Bulgaria, Turkey, and the Black Sea region as well as from Africa.” In order to develop irrigation works, and later to settle the Madeiran Islands during the early fifteenth century, large numbers of slaves were kid-

Enslavement of “Free” Persons 117 napped from the Canary Islands by the Portuguese. More were kidnapped to work the sugar plantations of the Azores and Cape Verde Islands, so that by the end of the fifteenth century the Canary Islands, estimated to have had a population of about one hundred thousand when first discovered in the fourteenth century, had been nearly decimated.” Piracy flared up again in the western Mediterranean with the expansion of the Ottoman Empire during the late sixteenth century, and for the next two hundred years Christians and Muslims captured one another, enslaving many of their captives. The North African states, especially Algiers and Morocco, came to depend heavily for manpower and external revenues on the “little war” of piracy conducted by the so-called Barbary pirates, especially during the eighteenth century.” But it was not only in Europe and the other slaveholding states of the Mediterranean that kidnapping was important. It was the major original and current means of enslavement in Southeast Asia, where its maritime version, piracy, was highly developed. Kidnapping (including piracy) was the second most important source of slaves in Burma and Thailand.” A substantial number of slaves in China, right up to modern times, were originally taken by pirates, especially on the Korean coast.’° And in Japan piracy was rampant and a primary source of slaves up to the Moramachi period.”’ In Africa the advanced slave systems established by the Arabs on the east coast, especially in Zanzibar and Kenya, relied entirely on organized kidnapping as the original means of enslavement and as the most important current means of enslavement during most of the nineteenth century.’* The same was true of the slave regimes established by the Portuguese and the Dutch in southern Africa.” All societies strongly forbade internal kidnapping and sale of free persons. In ancient and medieval Europe it was usually a capital offense; but the continuous enactment of laws against the practice indicates that it was never completely eliminated.*” In China, while it was also always a capital offense, the laws seem to have been far less effective than in Europe. This was particularly so in the period of the Northern dynasties (A.D. 386-618), when thousands of local citizens were reduced to slavery by semiautonomous war lords.*! Even more ruthlessly kidnapped were the aborigines of the borderlands “who had never been considered equals of the Chinese and would be arbitrarily enslaved.”*? It is this same feeling of ethnic distinctiveness that explains two other important instances where internal kidnapping constituted a major means of enslavement: the emirates of the Sudan and the indigenous semi-Islamized states of pre-European Malaysia. The Islamic emirs, sultans, and noble families who ruled the many petty states of northern Nigeria and other areas of

the Sudan not only raided the African pagans on their borders but frequently turned on the subjected tribes within their own state boundaries,

118 Slavery as an Institutional Process even when they were known to be converted Muslims. Raiding was so much a part of life for them and so essential to their wealth that they strongly resisted early British attempts to stamp out the practice. As one emir proudly put it: “Can you stop a cat from mousing?”® It is, of course, forbidden by Islamic law to enslave fellow Muslims, and technically it was also illegal to enslave members of the raiders’ own State, especially when such raiders were

sworn to protect the subjected tribes. However, the viciously predatory rulers of these unfortunate Africans easily got around both prohibitions. They either simply ignored them or justified the raids on the grounds that the captured groups were really pagans. And they circumvented the problem of raiding their own subjects by tacitly agreeing to raid across one another’s boundaries with the understanding that there would be no retaliation as long as the captives were Negroes.” The situation was even more blatant among several of the pre-European states of Malaysia, where there was not even an attempt to find excuses. In

Perak, for example, the raja made annual raids on his own villages and seized every nubile girl who took his fancy. Less organized raids involving both males and females were also common. Again, this outrageous disregard for their own subjects was due to racial, ethnic, and (in the case of the pagan tribes) religious differences between rulers and ruled.™ We come, finally, to the colonial slave regimes of the modern Europeans. In the case of the large-scale slave system of the Banda Islands established by the Dutch and the Perkenier-family descendants during the seventeenth and eighteenth centuries, there is no doubt whatever that kidnapping constituted the sole original and most important current means of enslavement. Even before the arrival of the Dutch, piratical kidnappings and enslavement of peoples were common in this part of the world. When the native population died out, the inhabitants of the neighboring chains of islands, especially Sangir, were raided for slaves; as the demand grew, people were kidnapped from as far away as Arakan on the west coast of Burma.*° The situation is more complex with respect to the more than 11.5 million Africans transported to the slave regimes of the Americas. Were they primarily prisoners of war or victims of kidnapping raids? Earlier I suggested that during the establishment of these slave systems most of the externally recruited slaves were prisoners of war. While the proportion of externally recruited slaves declined over the course of the late seventeenth and eigh-

teenth centuries, the fact remains that when the total number of Africans transported is considered in absolute terms, the vast majority came during the eighteenth and nineteenth centuries.*’ So we still need to ask whether they were mainly prisoners of war or kidnapped. To attempt an answer to this question we must take account of several factors: the various time periods over which the slave trade lasted; the areas of Africa from which the slaves came; and whether the so-called wars on the coast were fought essentially to acquire slaves or for other purposes.

Enslavement of “Free” Persons 119 To begin with the third factor, Philip Curtin has proposed that for the west coast of Africa during the eighteenth century most wars were waged deliberately in order to acquire slaves. The Senegambia region departed from the norm, in that purely political factors motivated warfare somewhat more frequently than economic ones.** Henry Gemery and Jan Hogendorn take much the same position, argu-

ing that the large-scale use of African slaves in the Americas was at least partly promoted by a highly elastic supply of slaves on the west African coast. They argue, further, that the low market price for slaves before the coming of the Europeans suggests that the taking of slaves at that time was incidental to warfare. As the demand increased during the eighteenth century, more and more wars were waged for the primary purpose of taking slaves.*?

We may now ask a crucial question: should we dignify the aggressive assaults and raids on neighboring peoples, waged exclusively for the purpose of acquiring slaves, with the term “warfare”? My answer is an unequivocal no. The more closely we examine these so-called wars, the more we come to realize that they were nothing more than sordid kidnapping expeditions incited by no other motive than the desire for the goods and money being offered by European traders and their agents.”” True, there were important

political consequences of many of these raids, especially in the Guinea coastal area, but these were strictly by-products of the organized kidnappings. In southwest Africa there were no such complications. For virtually the entire period of Portuguese devastation of this vast area, almost all Africans taken were captured in raids—most as in the rest of Africa conducted by African middlemen, but many by the Portuguese themselves.”! J. D. Fage has estimated that less than a third of all Africans taken over the entire course of the trade were kidnapped, while a little more than half were genuine prisoners of war.” I suggest instead the following conclusion, based on more recent scholarship.’ Of the 1.6 million Africans brought to the New World before the end of the sevénteenth century, as many as 60 percent may have been the captives of genuine warfare, while slightly less than a third were kidnapped. Of the estimated 7.4 million transported between 1701 and 1810, the proportions were reversed—that is, over 70 percent were kidnapped and under 20 percent were the victims of genuine wars.

Many of those kidnapped were taken in raids organized by the rulers of centralized polities such as Dahomey and Ashanti, which had advanced politically primarily as a result of the economic stimulus of slave raiding and trading.”* However, it seems highly probable that the majority were kidnapped in smaller, individually organized raids such as those described by the Efik trading chief, Antera Duke, in the region that is now the Calabar

province of Nigeria.”

The nineteenth-century picture is more complicated. Between 1811 and 1870 about 2.4 million slaves were brought to the New World, the vast ma-

120 Slavery as an Institutional Process jority going to Brazil and the Spanish Caribbean, especially Cuba. The early decades of the century were times of political turmoil in Guinea, so it can definitely be said that the majority of captives coming from this region were genuine prisoners of war.”° This was true, for example, of the large number of Yorubas who ended up on the expanding slave plantations of Cuba.”’

Most of the slaves who left Africa in the nineteenth century, however, wound up in Brazil (over 1.2 million of the total of about 2.4 million); we know the majority of these came from southern Africa and that almost all of them were kidnapped.”* We may conclude that for the nineteenth century a little over 60 percent of the slaves brought to the New World were kidnapped, while a little under 30 percent were genuine prisoners of war. In all, then, the overwhelming majority of slaves brought to ail regions of the New World were kidnapped persons, with no more than 30 percent being genuine prisoners of war. After 1700, prisoners of war outnumbered kidnapped persons in only a few regions during certain brief periods (Jamaica during the first quarter of the eighteenth century, Brazil during the first two decades of the nineteenth century, and Cuba during the first half of

the nineteenth century). | a

One final issue must be addressed in connection with enslavement by captivity in warfare and by kidnapping: it concerns sexual bias. A common view is that among more primitive peoples where slavery was not very important, there was a strong preference for women, but that with more advanced social systems and slave formations, the bias shifted toward the taking of male captives. The comparative data suggest otherwise. It is true that women were taken more frequently than men among small-scale, kin-based

peoples—although there are many exceptions. For example, among the Kerebe of Africa “male captives were as welcome as females,””’ and the Ibos apparently took men and women in equal numbers.'° Nonetheless, what Finley says of Homeric Greece holds true for most small-scale societies: “There was little ground, economic or moral, for sparing the lives of the defeated men. The heroes as a rule killed the males and carried off the females, regardless of rank.”'°’ It turns out, however, that this sexual bias in favor of women holds true for the great majority of peoples. There was certainly a decline in the tendency among more advanced peoples to kill off their male captives, especially when there was an economic need for slaves; but before the Atlantic slave trade we rarely find more males being captured than females, and the practice of massacring male captives remained prevalent even where they

were also enslaved. .

What determined sexual bias in the taking of captives was not the level of development of the society or the degree of structural dependence on slavery, but the use to which slaves were to be put (especially in the dominant mode of production), purely military considerations, and the problem

Enslavement of “Free” Persons 121 of security in the captor’s society. It is obvious that women and children were easier to take than men; they were also easier to keep and to absorb in the community. In addition, in most premodern societies women were highly productive laborers and, especially in Africa, they were frequently the main producers.'°” Even where men had traditionally monopolized the productive sector, slave women were absorbed as workers. After examining a hundred cases of the aftereffects of conquest in ancient Greece between the sixth and the second century B.c., Pierre Ducrey found that the practice of enslaving the women and children and killing the

men was no longer “normal” but was still quite common.'” The situation was not much different among the Romans. Indeed, some authorities suggest that their practices seem to have been closer to what we find among primitives. Mars M. Westington concluded his study of atrocities in Roman warfare by observing that “the slaughter of adult males and the enslavement of women and children is tersely mentioned with the regularity of a fugal theme.”'™ This is clearly an exaggeration, contradicted by some of Westington’s own evidence and later research; the male slaves on the latifundia had to come from somewhere. Nonetheless, the primitive practice of massacring the men and enslaving only the women and children was clearly attested in numerous instances.'” With the rise of the Islamic states we find a systematic effort to capture as many men as women in order to supplement the conquering armies of Islam

and reinforce their manpower.'’° Once these states were established, the age-old practice of favoring female over male captives returned. Among the

great majority of Islamic peoples after the ninth century, female captives and kidnapped persons fetched a higher price than males, even where slavery was economically important.'®’ Often radical shifts in sexual preference took place over time in certain societies, depending on changes in the demand for and uses of slaves. During the earliest periods of Mesopotamian and Egyptian history there was a decided preference for female prisoners, males being killed on the spot; later the bias moved in favor of males.'°® According to the Soviet classicist J. A.

Lencman, there was an important shift away from the tendency to take mainly female slaves between the world depicted in the /liad and that depicted in the Odyssey. After noting that female captives were mentioned 11 times as against male captives once in the //iad, in contrast with 46 references to female slaves and 34 references to males in the Odyssey, Lencman speculates that there was a higher incidence of kidnapping in the world of

Odysseus (pillage being conducted mainly to capture slaves and other booty), while the taking of slaves was largely a by-product of genuine warfare in the Iliad.'"’ I need hardly belabor the dangers of such evidence, if evidence it is. Lencman, however, makes a very good point in the course of this otherwise questionable speculation: a deliberate shift in favor of male

122 Slavery as an Institutional Process slaves strongly suggests that kidnapping is the major means of enslavement.

With this in mind, let us look at one area where there was a marked tendency toward such shifts in sexual bias in the selection of slaves, namely, the slaveholding and trading peoples of West Africa.

Such shifts are found there only after the contact with Europeans. The Aboh, who traditionally had taken women and children, with the coming of the Europeans took both sexes, keeping the women and children for themselves and selling the males to the Europeans.''° The Vai, before 1826, took only women and children. Between 1826 and 1850 they took mainly men, to

meet the demand on the coast. When the Atlantic trade dried up in about 1850, they returned to the practice of killing male captives and taking only

women and children.''' This changing sexual bias was even more pronounced among the Duala of West Africa, who until 1700 took mainly women and children, to meet their own traditional domestic needs; then between 1700 and 1807 shifted to an emphasis on males, to meet the needs of

the European traders; after 1807 returned to an emphasis on women and children when the export trade declined; then, with a shift in their own mode of production at the turn of the century, changed once again to the acquisition of mainly male captives, a pattern that continued until 1920, when slavery was finally abolished.''” What all this demonstrates beyond any reasonable doubt is that the supply of slaves was highly elastic on the west coast of Africa, even to the point

of being sex specific. The vast majority of Africans brought to the New World were not prisoners taken in wars either of their own making or of anyone else’s. As Equiano and other African ex-slaves who wrote their autobiographies so often insisted, the slaves were stolen from their homes by European-supported thieves.'!?

Tribute and Tax Payment Enslavement as a result of being part of tributary or tax payments is obviously related to warfare but should not be confused with either captivity or kidnapping. The vassal state may never have engaged in warfare with the

state to which it paid tribute; it may have voluntarily offered tribute as a means of preventing attacks or simply as a goodwill gesture to the more powerful state.

Most of the advanced premodern peoples who kept slaves on a large scale obtained some of them at some time by this means. It was unusual, however, for significant proportions of a slave population to be obtained as tribute, especially where slavery was of marked structural importance. Rome is a good case in point. The Roman economy by the period of the early empire was “basically a money economy,” as Richard Duncan-Jones points out, and tribute payments in cash were preferred (although payments

Enslavement of “Free” Persons 123 in corn were common).''* Tribute payment in slaves was important during only one period, the late second and first centuries B.c., when Roman tax farmers plundered the eastern provinces and took persons as slaves in such vast numbers “that when Nicomedes of Bithynia was asked for a contingent at the time of the Cimbrian wars he replied that the majority of his subjects had been carried off by the tax farmers and were now in slavery.”'!° The Islamic states and several of the advanced pagan states of Africa stand out as the slaveholding peoples who relied most heavily on tribute to establish and augment their slave populations. Although most of the slaves in the early Abbasid slave armies were bought, considerable numbers were obtained as tribute.'’® According to Ibn Xurdadhbih, between the years A.D. 826 and 828 two thousand captives of the Turkish tribe Guzz were sent from

the province of XurAsdn as part of their tax payment.''’ A considerable number of the elite slave corps found throughout the Islamic world came as

tribute, the most celebrated being the Ottoman janissaries, who were recruited by means of the devshirme (tribute of children from the Christian subjects of the empire).''® A large number of the public and private slaves of

Muslim Africa were similarly obtained.''? Among pagan states with advanced slave sectors Ashanti and Oyo were the two that relied most heavily on tribute. !”°

Where did the tributory slaves come from? Often they were persons who were already slaves in the tribute-paying state. Sometimes, however, the unfortunate slave was already part of a tribute paid by another vassal state in an international pecking order. Thus in the mid-nineteenth century the emir

of Adamawa (the primary slave center of the Fulani) was paid approximately five thousand slaves by his vassal states, of which he sent two thousand to the sultan of Sokoto.'*! Similarly Bornu, itself a tribute-paying state, received tribute in slaves from the Kwararafa kingdom, which in turn received slaves from its own vassal states.'’* Equally elaborate was the tribute system extending from the Guinea coastal state of Popo to the Ardra, who in turn paid to the Oyo, who were sometimes in vassalage to yet more powerful states.'?° Many vassal states, too weak to impose tribute on others, did not have enough slaves to meet their quota so were obliged to send their own “free” people as part of the tribute. Thus when Korea became a client state of the Mongols in the thirteenth century, most of the slaves sent as tribute were free persons.'* The same was true of many of the subjected peoples who paid tribute to the Aztecs.'”? But few societies have had a longer or more unhappy history as tribute payers than the Nubians. They provided slaves both from among themselves and from their southern neighbors as tribute to the viceroy of Egypt from the Nineteenth to the Twentieth dynasties, especially during the viceroyalty of Kush. Over two thousand years later the Nubians were still paying tribute in slaves to foreign conquerors. After the negotiated

124 Slavery as an Institutional Process truce with the Arabs in a.p. 651-652 (called the Bagt by Arab historians) a tribute in slaves was demanded by the Arabs. The terms are instructive. According to the Arab geographer Magrizi, it required that “each year you are to deliver 360 slaves which you will pay to the Iman of the Moslems from the finest slaves of your country, in whom there is no defect. [There are to be] both male and female. Among them [is to be] no decrepit old man or woman or any child who has not reached puberty. You are to deliver them to the Wali of Aswan.”'*® The insistence on slaves without defects touches on a problem that must have plagued all such payments. Indeed, wars were sometimes fought over the quality of the slaves sent as tribute. In the nineteenth century “the custom of the king of Bagirmi of sending his oldest, ugliest and most useless slaves to Wadai was one of the provocations which

moved the king of Wadai to attack him in 1870.”'!”’ | Debt Debt as a source of slaves must be examined with the greatest caution for, on the one hand, debt is usually a reflection of other causes such as poverty and, on the other hand, so-called debt-slavery has to be carefully distinguished from true slavery. Even so, it very often happens that persons get into debt not from poverty, but as a result of risks that were in no way pressing. While the distinction between slavery and debt-servitude is important, the fact remains that in all societies where debt-slavery existed, the possibility of the debt-slave falling into permanent slavery was always present. Debt, in short, may have been a direct or an indirect cause of slavery. Among the less commercially developed peoples, debt tended to create

slaves by indirect means. Throughout traditional Africa, for example, the practice of pawning was widespread. Usually it was not the debtor, but a member of his family, who was pawned or pledged as security for a loan. The Ashanti were typical. The pawn was usually a woman, often the niece of the debtor—his jural daughter in this matrilineal society. Most such pawns were restored to freedom, since most debts were repaid. Still, it occasionally happened that the debtor was unable to repay his loan on time and, given the high interest rate (50 percent a month according to one source!) got

deeper and deeper into debt to the point where he simply could not repay the debt. When this happened, the pawn became a genuine slave. '** Much the same situation prevailed among the tribal peoples and petty states of pre-European Malaysia, where debt may well have been the major source of slavery. The interest rate among the more primitive Bataks was a whopping 100 percent—“folding the debt” is the vivid local expression. Persons also fell into debt-servitude and later slavery, as a result of high bride-price and passion for gambling.'*’ The situation was not much better among the more centralized Islamic states. One notorious practice among

Enslavement of “Free” Persons = 125

the Arab rulers of these states, especially in Perak, was the imposition of spurious and heavy fines on the native population which the fined person had no means of paying. He would then go into debt-servitude to the raja and eventually, with the accumulation of heavy interest, fall into permanent slavery.'*° Debt as a direct cause of slavery was more common among the commercially more advanced peoples. It was one of the most important sources of slavery in ancient Mesopotamia,'’' and the second most important among

the ancient Hebrews,'** the premodern Koreans,'*? and the premodern

Thai.'**

The enslavement of persons for debt, or the sale of oneself or one’s relatives to repay a debt, was specifically forbidden among a wide range of peoples. Sometimes only the practice of “reducing” the pawn or debt-servant was forbidden. Among the nineteenth-century Damagaram of Zinder, Nigeria, “no individual could reduce another to slavery because of debt.”’'*> Islamic law forbade enslavement for debt although, as we have seen, the pro-

hibition was frequently circumvented. Debt-servitude was rampant in Hesiodic Greece and was the source of much unrest. In Athens one of Solon’s major reforms was to abolish the practice and, in so doing, reduce the possibility of citizens falling into genuine slavery as a result of destitution.'*° Many other Greek states followed Athens’ lead, but there were exceptions such as Gortyna in Crete. Further, in most Greek states a prisoner of war ransomed by a fellow citizen remained a slave to his ransomer until he repaid the debt. Citizens who did not pay their taxes could also be made public slaves. And in most Greek states metics, especially freedmen, could be reduced to slavery for debt. In the oriental regions of Hellenistic Greece the pre-Greek tradition of enslavement for debt continued after the Greek

conquest.'°” Roman law too was very harsh on debtors. In early Rome judgment-debtors might eventually fall into slavery, but as W. W. Buckland points out, “the position of iudicatus in early law is in some points obscure ...and the system was very early obsolete.”’** At all times, defaulting debtors were subject to compulsory labor.'°’ In several societies the prohibition of enslavement as a result of debt was part of a wider prohibition on the enslavement of free natives except for capital offenses. In both China and Vietnam the sale of one’s self or one’s wife or relatives into permanent servitude was forbidden.'*° However, the pawn-

ing and pledging of persons was permitted, and this provided an escape route that was frequently exploited. The sale of persons, as such, was not illegal: what was illegal was the failure to ensure that a sold person did not fall into perpetual servitude by making it clear to the purchaser that the person

being sold was “good” and not “base.” It is easy to see how such a law would frequently be broken. Court cases on the matter ended up in long ar-

guments, which sought to establish whether the purchaser bought the

126 Slavery as an Institutional Process “good” person honestly thinking he or she was “base” or whether the seller had taken sufficient care to make the status of the sold person known. '*!

| Punishment for Crimes | The enslavement of criminals who had committed capital offenses and other serious crimes was practiced in the great majority of premodern slave systems and in several European states down to the nineteenth century. Among a number of primitive peoples it ranked as the primary source of slaves—

usually only where slavery existed on a small scale. It was an important source, for example, among the Ibos of West Africa and the Goajiros of northern South America. Among more advanced premodern societies crime tended to be of less significance as a source of slaves. It was minor in the an-

cient Near East. In ancient Greece penal enslavement existed but was largely confined to metics, foreigners, and freedmen in central Greece; it was never a significant source of slaves. In Hellenistic Egypt it was of more economic importance; but since the main crime for which persons were penally enslaved was insolvency to the state, the difference between this source of slaves and enslavement for debt was slight.'*” In Rome penal slavery was a far more established institution: “a person convicted of crime and sentenced in one of certain ways suffered capitis deminutio maxima, and became a slave. It was essentially capital punishment, and the capitis deminutio had all its ordinary results.”'*’ Not all forms of capital punishment involved the reduction to penal slavery, and only some

categories of persons were affected (usually lower-class freemen). Only when the sentence was lifelong was it slavery, and a distinction was drawn between temporary penal servitude in the mines and permanent slavery. One variant in Rome harks back to the most primitive roots of slavery: persons who were condemned to die became penal slaves during the interval

between their sentence and their execution. This was particularly true of persons condemned to die ad gladium and ad bestias. Goods belonging to the servus poenae (penal slaves) technically had no owner, and only the emperor could manumit him or her. Such pardons, the so-called indulgentia generalis, were not uncommon, and although the pardoned person was restored to the status of freeman he had certain liabilities. It must be emphasized that in Rome the enslavement of criminals was essentially a penal matter: as a source of slaves it was insignificant. Penal slaves did perform economic roles—mainly in the mines—but their contribution to the Roman economy was slight.’ In several oriental societies penal slavery was a significant source of both public and private slaves. It provided the bulk of slaves among the ancient Vietnamese, for instance, although slavery was never of any real importance there.'*” In Korea, which had the most advanced slave system in the Ori-

Enslavement of “Free” Persons = 127

ent—and one of the most developed anywhere in the premodern world— penal slavery was never a major source of slaves.'*° It was of greater signifi-

cance in Japan. Here, prior to the sixth century A.D., the two primary sources were prisoners of war and the kinsmen of criminals (as well as the criminals themselves). However, as slavery gained in economic significance during the sixth and seventh centuries, these were replaced by poverty and destitution as the principal sources. '*’ In China penal enslavement was the foremost source of slaves. Strictly speaking, the enslavement of the families of condemned persons was the only recognized source of slaves in Chinese law. As we have seen, there were many other sources; but these were usually either illegal or extralegal. Significantly, those prisoners of war who were enslaved were first assimilated to the status of convicts. Unlike Rome, the strong emphasis on familial responsibility in China meant that a person’s wife and kinsmen were fully liable for his criminal actions. The number of such kinsmen varied but at times the law became draconian, involving the entire clan of a convicted person. Before the Han period the convicted person was always executed and his family enslaved. After this there was a growing tendency to enslave both the offender (in the case of the lesser capital offenses) and his family. Pulleyblank argues persuasively that penal slavery, being the origin of both public and private slavery in China, gave the institution its name, and influenced “Chi-

nese conceptions of the nature of slavery and ... the legal position of slaves.”'** The slave was always viewed as a criminal and, as such, base (chien) and subject to physical mutilation. This is why the sale of “good” persons into slavery was so abhorred in Chinese law. It also explains why kidnappers of free persons immediately branded them. This was the surest sign of their criminal origins and made them much easier to sell. C. M. Wilbur argues that a distinction was always made in Chinese law between convicts and slaves; Pulleyblank, while agreeing that the distinction existed in law, argues that in practice it did not. China is unusual too in that penal slaves often ended up in the hands of private owners. Usually presented as gifts, they were sometimes simply seized by unscrupulous officials and officers. As in Rome, only the emperor could manumit penal slaves; but the issue was complicated because so many were in private hands. Apparently the manumission of penal slaves by private owners was condoned though never recognized in law.'”’ J. Thorsten Sellin, drawing on previous works, shows that there was a threefold relationship between slavery and the penal system in the history of Europe.'”° First, slavery remained a form of punishment throughout the Middle Ages, though penal slaves did not serve as an important source of slaves even where the institution remained important. Second, over the medieval centuries the nature of punishment of free persons was strongly influenced by the kind of punishment originally inflicted only on slaves. Slav-

128 Slavery as an Institutional Process ery, in other words, had an increasingly retrogressive effect on the treatment of convicted persons. In the words of Gustav Radbruch: “To this day, the criminal law bears the traits of its origin in slave punishments . . . to be punished means to be treated like a slave. That was symbolically underscored in olden times when to flogging was joined the shaving of the head, because the

shorn head was the mark of the slave ... Slavish treatment meant ... not just a social but a moral degradation. ‘Baseness’ is thus simultaneously and inseparably a social, moral, and even aesthetic value judgment . . . the diminution of honor, which ineradicably inheres in punishment to this day, derives from slave punishments.” "”! There was also a third relationship. Penal slavery became from the late Middle Ages down to the nineteenth century a means of recruiting labor for the mines, the galleys, and other public works, especially in Spain, France, Italy, and Russia. In quantitative terms, the number of such slaves was never great among the Western Europeans;'*’ in Russia, however, from the late seventeenth century on, it was one of the major sources of slaves in what was a very large-scale slave system (even if one excludes the so-called serfs, most

of whom were genuine slaves). Punishment for crimes was the source of nearly all the vast number of public slaves who worked in the mines and developed the Siberian hinterland.'”’ Penal slavery was prohibited in Islamic law. Indeed, the introduction of Islam to a country usually terminated this means of enslavement.

Slavery was often a punishment for capital offenses, whatever these might be. Typical of the list of persons considered criminal, and hence enslaved, in a primitive society is that of the Ibos. The roster included adulterers, those who sold or rented communal property, “quarrelsome person(s] whose headstrong activities might lead to war,” very disobedient children, thieves, sorcerers, those accused of witchcraft, and any seizable members of a divorced woman’s family where her bride-price had not been returned.'”* The list is usually much shorter among more advanced peoples: those who committed treason and particularly horrible crimes such as patricide (in the case of Rome), and (in Greece) persistently promiscuous daughters, metics who enrolled as citizens or refused to pay the metic tax, freedmen who failed to live up to their obligations to their patrons, and foreigners who married Athenian women.'*° In some societies unscrupulous rulers were occasionally tempted to increase the number of crimes for which persons might be executed or enslaved. In West Africa the list grew with the expansion of the Atlantic slave trade. A number of Africans who ended up on the shores of the Americas were tricked into slavery. A common practice was for several of the many wives of an unscrupulous chief to seduce unwary young men, then accuse them of committing the capital offense of adultery with the wife of the chief. The practice became known among the Sherbro as “woman damage.”'°° Rarely was it more than a minor source of slaves.'>’ In early modern

Enslavement of “Free” Persons 129 Europe we find a similar increase in the number of crimes punishable by enslavement in the galleys.'”®

Abandonment and Sale of Children This was a widespread source of slaves but apart from the ancient Mediterranean it was rarely a major one. Among kin-based preliterate peoples it was often due to poverty; just as frequently it resulted from some attribute of the child, such as birth defects to which taboos were attached. Twins, slightly deformed children, those born with peculiar birthmarks, breech babies, and the like might be either killed or exposed. The only unambiguous case I know in which exposure of infants became the major source of slaves was among the Fulani of Borgou in northern Benin, Nigeria, during the nineteenth century. The Batomba of Borgou, who did not practice slavery to any significant degree, had the custom of killing any child whose first tooth appeared in the upper jaw because such a child was felt to bring “disaster, illness, and death to his family.’ Occasionally a child was spared if a certain tribal official, the gossiko, was available and willing to raise it, but the vast majority were left to die from exposure. When the immigrant Fulani arrived, they eagerly picked up these children and reared them as slaves. In time the Batomba hosts would simply hand the children over to their Fulani guests, who used them not only as shepherds and domestics but to perform the agricultural duties they themselves despised.'”” While never a major source, the exposure and sale of infants was not unimportant in the ancient Near East and the Orient. In China the vast majority of such infants were girls, reflecting not only the strong cultural preference for boys (in this respect the practice is closely associated with female infanticide), but also the greater demand for female slaves mainly for domestic and sexual purposes. Most of these girls eventually ended up as concubines or prostitutes. This form of slavery continued in China right down to the twentieth century, where it was disguised as a form of adoption—the

so-called Mui-tsai institution.'°° |

Poverty was the main reason for sale and exposure in the ancient Near East, and during periods of hardship it became a not unimportant source of slaves. The practice was referred to as being “placed in a pit” or “thrown to the mouth of a dog,” terms that grimly indicate the outcome if the child was not fortunate enough to have been enslaved.'®! It is in the ancient Mediterranean and adjacent lands that child exposure became a significant source of slaves. Although forbidden in one or two of the Greek city-states (Thebes, for example), it was permitted in most and was a common though never major source during classical times. It was of more significance in the Hellenistic world and in Asia Minor, and was to 1ncrease in significance during the period of Roman rule. In Rome itself the

130 ~=Slavery as an Institutional Process

practice was negligible during republican times but became a primary source of slaves from the period of early empire on. William V. Harris claims that “no other source within the empire can have made a major contribution to filling the gap left by slave-born slaves.”'°* Poverty was not the only reason for exposure. In Rome, as in other parts of the world, persons from all classes exposed unwanted children, again with girls being exposed at a disproportionately higher rate.'™ Both Babylonian and Roman law remained ambiguous on the issue of exposed children, especially concerning the restoration of their status as free persons. The only section of the Babylonian code that applied to child exposure, despite its frequency, was the edicts on adoption. From these it has been inferred that the child could be taken back by its parents at any time before it became an adult, but not thereafter.'“* Roman law, in keeping with its basic rule that free persons could not be reduced to slavery except for capital offenses, maintained that the foundling who was saved and brought up as a slave could reclaim her or his freedom if free birth could be proven. The burden of proof, however, was on the slave, and since it was virtually impossible to secure such proof the law really amounted to very little. What is more, in the Greek-speaking part of the Roman empire it was required that a ransom be paid for the cost of upbringing, should such a person be lucky enough to prove his free birth. (I should add that the Roman emperors consistently refused to recognize the legality of this rule.)'®

Self-Enslavement Poverty was, of course, one of the main reasons for self-sale, and we have already noted that in several advanced societies such as China and Japan it was at times a major source of slaves. In Russia between the seventeenth and nineteenth centuries self-sale as a result of poverty was the most important reason for enslavement among the mass of domestic slaves. Richard Hellie goes so far as to call Russian (private) slavery a welfare system.'© Yet there were reasons other than poverty why persons sold themselves.

Sometimes it was because of political rather than economic insecurity. Strangers who found themselves cut off from their kinsmen in tribal societies often sought self-sale into slavery as the only path to survival. This was quite common, for example, in the Kongo during the unsettled years of the nine-

teenth century.'°’ In primitive Germany it was often the only means whereby isolated persons could procure land and protection.’®® Another cause of self-enslavement was the sale of self and relatives in order to escape

either military services or prohibitive taxes—whether in cash, kind, or corvée labor. This was not infrequent in China, especially during oppressive regimes but it was the most marked in Korea. Indeed, during the Yi dynasty (A.D. | 392-1910) self-enslavement was the single most important source of slaves.

: Enslavement of “Free” Persons 131 There were a few peoples among whom individuals might sell their relatives, wards, and unwanted wives and children not out of economic necessity but for pure economic gain. The Ainu, for example, were notorious among the neighboring Siberian tribes for this practice.'’” We must, however, view

such claims with great caution, for they often turn out to be self-serving propaganda among the slave-purchasing and slave-trading groups. GrecoRoman talk had it that the peoples of Asia Minor eagerly sold their wives and relatives into slavery out of sheer greed, but there is no proof of this.'”' The Slavs suffered the same stereotype during medieval times.'’* And one must treat with equal caution the claim of Islamic traders that the nomadic peoples of Asia, especially the Turks, avidly sought out the traders in order to sell their relatives and fellow tribesmen.'”’ Finally, enslavement resulting from marriage to a slave must be viewed as a form of self-enslavement. In many primitive societies, as well as in most of the oriental and European slaveholding societies from ancient to modern times, we find this as a minor form of enslavement.'”* In most cases it was

free women marrying slaves, who paid the penalty of their freedom. Even where the laws prohibited it, men usually found ways of marrying slave women with impunity.

|

Enslavement by

Birth

|

|

ENSLAVEMENT BY BIRTH was, naturally, the consequence

of earlier forms of enslavement, but in all societies where the institution acquired more than marginal significance and persisted for more than a couple of generations, birth became the single most important source of slaves. Of the great majority of slaveholding societies the stronger claim may be made that birth during most periods was the source of most slaves. The discussion of other means of enslavement has implied estimates of the relative significance of birth, which will not be repeated. However, in view of certain common misconceptions concerning the capacity of slaves to reproduce themselves, a few crucial remarks may clear up some popular confusions. First, it is essential to distinguish between the biological and the social reproduction of a slave population. By “biological” I refer to the ca-

pacity of a slave population to produce a number of persons equal to or greater than itself, whatever the status of the succeeding generation. The only issue is whether the total number of deaths is balanced or exceeded by the total number of births. By “social” reproduction of a slave population I refer to the degree to which it is able to reproduce itself when, in addition to birth and death, nonnatural factors are taken into account, the most important being manumission and the immigration/emigration rates. As we shall see in Chapter 10, most slave populations had high manu132

Enslavement by Birth — 133

mission rates. One major consequence was that many slave populations that were biologically self-reproductive were nonetheless socially nonreproductive, because of the social leakage of persons from slave to “free” status. This was the case with many (perhaps most) Islamic slave societies and with most of the Spanish-American slave societies during the eighteenth century. The Mexican and Peruvian populations, for example, virtually disappeared by the end of the eighteenth century not because they were not biologically reproductive, but because of the social loss due to manumission.' A second important point is that where there is a large influx of externally acquired slaves, the claim that a slave population is nonreproductive may well be based on a demographic illusion. In such situations the excess of deaths over births may be entirely a function of the abnormal age structure of the population brought about by the large number of adults—especially adult males—in the population. It is not incorrect, but it is certainly misleading, to say that such populations fail to reproduce themselves naturally. Age-specific mortality and fertility rates may be quite normal. This was true, for instance, of the Cuban and English-speaking slave populations during the nineteenth century.” The failure to distinguish between age-specific and general rates of birth and death has led to unwarranted generalizations about slave populations’ failing to reproduce out of despair with their lot. True, there have been a few such cases but they are rare in the annals of human slavery. The instinct to reproduce usually triumphs over despair, so that the exceptional cases become all the more poignant. Which were these exceptional cases? The one unambiguous instance is Jamaica during the second half of the eighteenth century. Here all the available data suggest that not only was the mortality rate abnormally high but, more extraordinarily,

slave women absolutely refused to reproduce—partly out of despair and outrage, as a form of gynecological revolt against the system, and to a lesser extent because of peculiar lactation practices.> The other exceptional cases we can only guess at, given the poor quality of the available evidence. From

tury.” |

the ancient world the slave population of rural Rome during the last two centuries of the republic seems likely,* as from the modern world does the slave population of the coffee region of Brazil during the nineteenth cenA third point to be stressed is that even if a slave population is biologi-

cally nonreproductive, birth may still remain the single most important source of slaves. There is a tendency among historians to leap from the (correct) observation that birth failed to meet the total demand for slaves to the (often incorrect) assertion that other factors were more important. The simple mathematics of reproduction contradict this thesis. The traditional view of historians of ancient Rome that “most of the need for slaves” was met by birth after the completion of the aggressive wars in the early empire has recently been sharply disputed.° I do not have enough information on ancient

134 Slavery as an Institutional Process i | Rome to argue the issue in meaningful statistical terms, but the experience of Jamaica during the eighteenth century may be instructive. We have already observed that the Jamaican slave population during most of the eighteenth century was unusual for its biological and social nonreproductivity. Between the end of the seventeenth century and the middle of the eighteenth, the enormous growth of the slave population was due to the massive importation of slaves from Africa. Males outnumbered females to a degree greater than any estimate ever suggested for the slave population of imperial Rome. And yet by the end of the 1760s Creole slaves outnumbered Africans. In other words, in spite of a demographic environment significantly worse than that of Rome during the period of the early empire, birth remained the most important source of slaves.’ Thus the comparative data strongly support the traditional view expressed by historians such as W. W. Buckland and R. H. Barrow that birth was “in historic times, by far the most important of the causes of slavery.” Let us turn now to the more important factors that influence the social reproduction of slave populations, specifically the social and legal patterns affecting the inheritance of slave status. The manner in which birth determined status was exceedingly complex, varying across cultures as well as within the same society over time. What complicates the issue is the fact that in all slave societies in which the number of slaves was of any significance, free persons interbred with slaves, thus making it difficult to determine the status of the offspring of mixed parentage. Sometimes these were free, sometimes slave; sometimes they occupied an intermediate status, depending on the sex, status, and power of the free parent as well as the relationship between father and mother. A consideration of the factors determining slave status at birth cannot be divorced from those determining free status, since once slavery was established not all persons born free were necessarily the children of parents both of whom were free. There were five ways in which slave status was determined by birth: (1) by the mother only, regardless of the father’s status; (2) by the father only, regardless of the mother’s status; (3) by the mother or the father, whoever had the higher status; (4) by the mother or the father, whoever had the lower status; and (5) by neither, the child always being free regardless of the status of either or both parents. The last case, of course, refers to incipient (nonhereditary) slavery and is not, strictly speaking, genuine slavery as we understand and use the term. Such cases are, however, important in any attempt to comprehend the origins of slavery. Next we observe that where both parents were free, there were several possibilities for determining the status of the child, for a category of free persons may be determined by birth through, or inherited from, the mother only (matrilineal societies); the father only (patrilineal societies); both parents (double unilineal and bilateral); or optionally from either parent, that of the one with the more favorable status being stressed.

Enslavement by Birth 135 When we consider these two sets of rules of inheritance of status—one determining slave/free status and one determining categories of free status—a useful distinction emerges among slaveholding societies: there was

one group, the majority, in which the rules determining the inheritance of status for the children of parents both of whom were free differed from those determining the inheritance of slave/free status, while in the second group the rules of status determination were the same whether slave/free status or category of free status was being determined. In other words, there were some slaveholding societies in which, say, father’s status deter-

mined category of free status but slave/free status was determined by mother’s status, while in others the fact that father’s status determined cate-

gory of free status among the children of free parents also meant that father’s status determined child’s status even if the mother was a slave. A mo-

ment’s reflection will indicate that where there was a high incidence of mixed unions (which was true of the vast majority of slaveholding peoples), it made a huge difference to the progeny of such unions and to the number of slaves recruited by birth whether the rules of determination were similar

or different. ,

A careful examination of the comparative data reveals that all known

slaveholding societies fall, empirically, into seven main groups when classified in terms of the joint operation of the rules determining categories of free status for the children of free parents and categories for the children of parents one or both of whom were slaves. I have followed the convention of kinship studies in designating each such class by the name of the society that

was most typical of it: Ashanti, Somali, Tuareg, Roman, Chinese, Near Eastern, and Sherbro. In the discussion that follows I concentrate on the “ideal type,” so to speak, and conclude with examples that conform to that type. The fact that societies belong to the same type in no way implies any historical link between them.

The Ashanti Pattern | The rule of status inheritance for children of nonslave parents in this group was unilineal, uterine (basically matrilineal), and the same for the children of parents who were either mixed (one slave, one free) or were both slave. The Ashanti provide the classic case. The children of the union of two slaves belonging to the same master remained slaves of the parents’ master. The children of the union of two slaves of different masters also remained slaves, but belonged to the master of the mother. Among the Ashanti there was a special name for the children of two slaves, nnonokfo mienu mma or some-

times afono mma. The children of the union of a male slave and a free woman were always free and were in no way under the potestas of their slave father or his master. The children of the union of a free man and a female slave were always slaves and were known as Kanifa, or half-Akan. Al-

136 Slavery as an Institutional Process though they grew up to consider themselves Ashanti, they were still slaves unless they were formally adopted. The child whose father was the master of his mother was of special importance to the matrilineal Ashanti, for only over such children did the biological father have complete authority (the potestas of the natural children of a man by his nonslave wife were held by his wife’s brother). According to R. S. Rattray, “the children of these children in the female line were considered slaves forever.”” The need for such children was an important reason for slavery in many matrilineal societies. The Ashanti pattern was the direct consequence of structural principles typical of matrilineal! societies. Hence it was found not only among all other Akan-speaking groups but, with the notable exceptions of the Tuareg of the Sahara and the Merina of Madagascar, in all matrilineal slaveholding societies. To cite some of the more important cases, the pattern was found among the Mende, who claimed the children of their female slaves on the proverbially expressed grounds that “mine is the calf that is born of my cow.” Although far more ruthless than the Ashanti in claiming the children of their female slaves, sometimes to the point of disrupting their marital unions, Mende masters tended to manumit their slave concubines, and as such, their children by them, to a greater degree.' We find the same pattern among the Imbangala of Angola, of whom J. C. Miller writes: “Because one’s primary social standing derived from the mother under Imbangala rules of matrilineal descent, the child of mubika [slave] mother (whatever the status of the

father) assumed the mubika status.”'' It is the same with the important Slaveholding nineteenth-century Kongo group, the Mbanza Manteke, among whom permanent slave lineages were produced from the progeny of female slaves.'? The slaveholding Indian tribes of the northwest coast of America are of special interest because, while sharing strikingly similar cultural patterns, they differed in terms of the rule of status inheritance among nonslaves. The more northern tribes—the Tlingit, Haida, and Tsimshian—inherited status matrilineally, and here we find the Ashanti pattern with respect to the children of slaves and of mixed parentage. Rank distinctions were even more marked in these societies than among the Akan, so that even the children of a chief and a female slave could not expect ever to be manumitted. Indeed, a highly ranked person who married a lowly ranked “‘free” person ran the risk of being killed.'” Under these circumstances it was extremely rare for a free woman to bear children for a slave. Both woman and child were disgraced in

the unlikely event that this happened.'* Arnong the Bella Coola Indians mixed unions in which the mother was the nonslave were less uncommon. In such cases the kinsmen of the woman always bought and freed the slave father of her child and both during pregnancy and after there were elaborate purification ceremonies and gift giving undertaken to wash away the “stain” of slavery from the child. Significantly, if the father was the nonslave and

Enslavement by Birth 137 wealthy, similar ceremonies might be performed should he wish to elevate the status of his child, but the status of such children was always precarious no matter how powerful and wealthy their fathers.'° Among the Goajiro of South America the pattern was strikingly Ashanti, although it had elaborations of its own. Nonslave women always produced nonslave children, regardless of the status of the father. Sometimes, however, the master might pay the bride-price for a free woman to marry one of his slaves. Even when this was done the children still belonged to the mother and her kin group. The master of the slave husband was compensated for his bride-price payment by being given the bride-price of the first daughter of the union.'®

The Somali Pattern In this group status was determined by the father (whatever the status of the mother), and when both parents were slaves, the ownership of the child still was determined by the slave father. The Migiurtini Somali are the prototype of this relatively unusual pattern. Among them the child of the union of two slaves remained a slave but, unlike the Ashanti, the child always belonged to the master of the father. The master of the slave husband paid a nuptial gift to the master of the slave bride, and this legally compensated for the owner-

ship of the children of the slave woman.'’ The children of the union of a slave woman and a free man were free if the father acknowledged paternity.'° Unions between slave males and “free” women were so strongly prohibited that they were nearly unthinkable. However, there was a low-caste, nonslave group among the Somali, the Sab, with whom male slaves were al-

lowed to intermarry, and here the same rules applied: the children of a female Sab and a male slave were slaves, while those of a male Sab and a female slave became Sab.'” The rigidly patrilineal principle of the Migiurtini Somali was rather un-

usual among slaveholding groups. The Margi of northern Nigeria, who today live in the states of Gondola and Bornu, were another case in point. James H. Vaughan tells us that in this patrilineal society “the status of the mafa [slave] is hereditary in the male line.”?° Most slaves were women. There was, however, an important office filled by a person known as the birma, a kind of acolyte to the king, which was always held by male slaves.

Birmas were permitted to marry free women and, strictly speaking, their children were slaves of the king. However, in the case studies detailed by Vaughan it turns out that both the king and one of his sons who married daughters of their birma paid the latter bridewealth. Vaughan could get no satisfactory answer from his informants to this “seeming anomaly.”*! Some Ibo groups among whom slavery was a well-established institution

came close to the Somali pattern. Among most tribes marriage between

138 Slavery as an Institutional Process

slave and free of whatever sex was strictly forbidden. However, in the Okigwi area of Ndizogu it was possible for slave men to marry free women as long as the latter were from some remote village. The parents of the bride were compensated for the loss of status by the much higher bride-price they received for their daughter in such marriages. The children of such unions became slaves of the father’s master. Among other Ibo groups if a married woman had a child by a slave lover, the child inherited the status of his legal father—the woman’s free husband—and was therefore free.”

During certain periods and certain areas of the ancient and modern world there were cases that may be classed as Somali. In Homeric Greece “‘it was the father’s status that was determinative” for both free and slave.’* In certain of the Greek states this pattern held throughout the ages of antiquity. Indeed, status inheritance seems to have become increasingly Somali in an-

cient Crete. According to J. Walter Jones:* ,

The children are free or slaves according as to the spouse, whose home it is, is free or slave, so that if a free man goes to live with a slave wife the children are slaves of the master in whose household the mother lives. Later, how-

ever, children of a free man by a slave wife were always free in Crete whether he went to live with her or not.

In the modern Americas we find a Somali pattern in Maryland between 1664 and 1681, where the father determined the status of the child for both free persons and slaves. Although the Roman rule was adopted in 1681, the courts of Maryland were inclined to ignore it in the unusual cases where the

rule worked on behalf of the slaves.” Although it is nowhere explicitly stated, the same situation may have prevailed in early Virginia. It was not until 1662 that an act was passed that defined the status of the child according to that of the mother. According to A. Leon Higginbotham: “Prior to passage of this statute it had been an open question as to whether the normal doctrine of English law would be applicable—that the status of a child would be dependent on the status of the child’s father,” and since a similar act was not passed in New York until 1706 it may be assumed that at least in

some cases the Somali rule applied.*° During the seventeenth century the Somali rule held in the French Antilles up to 1681, and in South Africa for a few years after that.’

The Tuareg Pattern The Tuareg of the Sahara stood, literally, in a class by themselves. They were doubly idiosyncratic in that, first, they were the only Islamic people with a matrilineal pattern of status inheritance for the children of “free” parents, and second, they were the only matrilineal people among whom the status of the children of mixed parents was patrilineally determined. Both

Enslavement by Birth 139 peculiarities are directly attributable to the major role of slavery in the socioeconomic life of this group. Following Islamic practice, the children of a master and his slave concubine were “free,” and if a Tuareg noble married the slave of another man “he could claim nobility for the children by paying a high bridewealth to the original owner.”** Pierre Bonte, who worked among the Kel Gress group, tells us that this frequently happened, “the children generally having the status and rights of the father.””’ The assumption of full hereditary rights on

the part of the children of slave wives, however, was peculiar to the Kel Gress and did not hold for the Kel Ahaggar group. It should be noted that this kind of status inheritance would have been quite impossible among the Ashanti or any other matrilineal slaveholding group. More in keeping with the matrilineal principle is the fact that the children of two slaves belonged to the master of the female slave, although the master of the male slave paid a bride-price to the master of the female and became her fictive father-inlaw.°°

In summary, the pattern is one in which status was inherited matrilineally by the progeny of intracaste unions but was inherited patrilineally among those of intercaste unions. Murdock is certainly correct in his hypothesis that this peculiar arrangement was developed to preserve the purity of race and to prevent the transmission of proprietary powers to the substantial number of persons of mixed blood, and hence of slave/free ancestry.2!

The Roman Pattern This was the classic pattern found in many highly developed slaveholding systems and in most of the Western slave societies. The rule of status inheritance differed for the children of free and of mixed parentage, and cus-

todial powers differed with respect to the children of parents who were both free and those whose parents were both slaves. W. W. Buckland describes the Roman case with his customary clarity: “The general principle is simple. The child born of female slave is a slave, whatever be the status of the father; if the mother is free the child is free, whatever the status of the father.”°* The law further states that “the slave issue belongs to the owners of the mother at the time of the birth, not at the time of conception.” Roman law provided a rationale for this rule. According to Gaius, it was the rule of the ius gentium: where there was no connubium, the child took the status of

the mother.” ,

The Roman pattern should not be confused with the Ashanti, in spite of some superficial resemblance. The Ashanti pattern had a single principle of status inheritance for the children of both free and mixed parents, whereas the Roman pattern used different principles: patrilineal for the free, matrilineal for the slave. Emanating from this was another crucial difference: free

140 Slavery as an Institutional Process women among the Ashanti did not lose rank or the right to determine the status of their children if they produced children by slaves. As with all the patterns, exceptional situations did develop. There were a few cases where free women could give birth to slaves: (1) the child of a free woman who cohabited with a slave in agreement with the latter’s owner that any issue would be the slave of the owner, and (2) the child of a free woman who cohabited with a slave knowing him to be one. Hadrian abolished the first rule but, oddly, not the second. The apparent anomaly is explained by

the fact that the second law was more concerned with the proscription of unions between free women and slaves. Cohabiting free women became slaves, and the law grew progressively harsher on the issue. Their children became slaves according to the general rule.** There were also cases where the child of an ancilla (female slave) was free. Among other exceptional cases, if the mother was free anytime between conception and birth, and was legally married, the child was free; but if the pregnancy was volgo conceptio, the child became a slave.” The general rule was that the law operated in the

interest of the unborn child. Thus if the mother was a slave at the time of birth but was already entitled to manumission, with the delay in her emancipation resulting from administrative factors, the child would be free. The Roman pattern of status inheritance was widespread in both the Western world and parts of the oriental world. Most Western societies, whether influenced by Roman law or not, had this pattern, and it was characteristic of many Indo-European slaveholding peoples. In ancient India up to the period of the Buddha, the Roman rule prevailed and was so strongly maintained that even the son of a king by a slave concubine could have been

reduced to slavery. The rule, however, changed drastically in Mauryan times.*°

The Athenian experience conformed in general to the Roman pattern but, as Glenn R. Morrow observes, it is a “little more difficult to determine and seems to have varied in different periods.”*’ The rule was apparently Roman during the fifth and fourth and later centuries B.c., although there is reason to believe that in certain special cases children of slave mothers and free fathers were freed. Significantly, Plato took a harsher stand on this issue and would have had the children of mixed parentage inherit the status of the slave parent.”®

The Roman pattern held among many central and northern Germanic peoples, who applied it with extreme rigidity. Early Norse law punished cohabitation with a slave by a free person of either sex with enslavement to the master of the slave.°” During later periods, the laws were relaxed and a free woman who bore a child for a slave could pass on her free status to the child, whereas a free man, even a master of the mother, could not do so except under unusual circumstances.*” The pattern changed again in Sweden at the end of the twelfth century, as we shall see. “Thrall born of thrall was thrall”

Enslavement by Birth 141 went the Icelandic saying.*' In Icelandic law the child of a free woman and a

thrall was free but could not inherit. A man had to emancipate his thrall concubine if their child was to be free. Should the child be conceived before the mother was manumitted, the child was free but declared a bastard and

could not inherit.*” |

Much the same pattern seems to have existed among the Celts, although the legal texts provide little information on the subject. From the law covering the relation between free Welsh and servile foreigners we deduce that the situation was of the Roman type. We know from the Book of Iorwerth that children of free Welsh women and servile foreigners could be given the free status of the mother if the mother so chose. Ironically, it appears that this status was only given the child when the servile foreign father rejected his paternity. Servile foreigners were sometimes quite prosperous, which explains both the pregnancy and the woman’s frequent preference for penurial status for her child.** The child of a servile foreign woman and a free Welsh father, according to the same laws, took the status of the mother. Not all the Germanic and Celtic peoples followed the Roman rule. We have little information on the tribal Irish, but the ancestral myth of the most famous of the Irish royal families—Ui Neill—may have some significance. The reputed ancestor of the family, Niall Noigiallach (Niall of the nine hostages)—who, legend has it, lived in the early fifth century—was the son of Eochu Mugmedon (“lord of slaves”) and a British slave girl. This definitely suggests a non-Roman, possibly a Somali, pattern of status inheritance, at least among the royal families.” _ Several Germanic groups that rose to power in the wake of the Roman empire also had non-Roman patterns, especially the Lombards and Visigoths. So did Christian Spain from the thirteenth century on. At the same time, the Roman pattern was by no means peculiar to the Indo-European peoples. During the height of Japanese slavery we find not only the general rule that the children of a slave woman were slaves belonging to her master regardless of the status of the father, but, strikingly, a regular exception to the rule which held that the children of a free woman who cohabited with a slave were the slaves of the master of the father.*°

| The Chinese Pattern The Chinese pattern contrasts in many striking ways with the Roman. As in the Roman model, the rule differed for the children of free parents and those of mixed parents, but in almost every other respect the models were unlike. Niida Noburu has discussed the issue at length, deliberately highlighting the contrasts with the Roman pattern.*’ From the period of the Han dynasties on, the basic Chinese rule was that in cases where the parents were of different statuses, the offspring of the union became a slave. In other words, the

142 Slavery as an Institutional Process principle of deterior condicio operated: the child always took the status of the parent with the lower status. Between the eighth and twelfth centuries A.D. if a freewoman had a child by a slave, the child became a state slave and the woman was severely punished. However, in the event that a commoner girl

had married a slave in the genuinely mistaken belief that he was “good” (that is, not a slave), the child remained free. During the Yuan period (roughly the thirteenth to fourteenth centuries A.D.) slaves and “good” persons were allowed to intermarry, but the wife always assumed the status of

the husband—that is, slave women married to “free’> men became free, while “free’’> women married to slaves became slaves. What this amounted to was an exception to the basic rule, in that the issue of such marriages followed the status of the father. There was, however, a regular exception to this exception. If a slave was married to a freeman and had an illegitimate child, the child followed the status of the mother, whereas if the mother was the free partner and had an illegitimate child, the child remained free. It is interesting to note that the legal logic behind both the exception and the exception to the exception in Chinese law was identical to the logic of the general Roman rule of the ius gentium, which stated that where there was connubium the child took the status of the father and where there was none it took that of the mother. Because the Chinese of this period allowed intermarriage, the rule had markedly different social consequences. Though late in the development of slavery, the laws of Yuan can be interpreted as an extension and humanization of earlier Chinese practice. Prior to the thirteenth century A.D., women were always strongly prohibited from having relations with slaves and, as we have seen, the children of such illegal unions became slaves. The practice, however, did take place and the reforms of the thirteenth century simply gave it legal sanction. The woman and her child still paid a penalty in that both continued to be reduced to slavery. The major differences were the privatization of the penalty (the child now became the slave of the father’s master) and the humanization of the sanction: instead of

punishing the slave and banishing the woman, the lovers and their child were allowed to form a recognized union (though one that would produce permanently “base” or slave issue). The new regulations of Yuan were also an extension and humanization of the far more frequent unions in the opposite direction: those where “‘free’’ males entered into concubinage with slave women. Concubinage was long recognized in the law, and frequently the children of such unions were emancipated.** The new rules simply elevated such unions to the status of full marriages, the children of which became fully legitimate heirs. The Chinese case nicely illustrates why it is that we should not rely exclusively on law in interpreting the practices of a people. The basic rule of deterior condicio—that the child always takes the status of the lower parent—-strikes one as harsh, and if practice had conformed strictly to law there

Enslavement by Birth 143 would have been very few freed slaves in Chinese history. Practice, however, never did conform exactly to the dictates of law; in the end, it was law that

during the thirteenth and fourteenth centuries finally gave way to practice by legitimizing both exception and exception to the exception. | In practical terms Wang Yi-T’ung seems correct in arguing that over the full course of Chinese history, “the status of slave progeny is rather illdefined and consequently permits a high degree of fluidity in Chinese slave institutions.””” In fact, it was this unusual fluidity that partially accounts for the persistently small proportion of slaves in China in spite of the vast numbers of persons enslaved: “This lack of genealogical force sharply distinguishes Chinese slavery from the conventional consummate pattern [and] ... was one of the chief factors which prevented China from becoming static, and it may have given comfort and hope to those who aspired to freedom and higher social activity.”°° This statement, true of China, could apply equally to any number of slaveholding societies such as Rome and almost all

of the Islamic world. |

What happened where there were no humanizing exceptions to the basic Chinese rule either in law or practice? The cases in point are most instruc-

tive. We expect that if such a severe restriction of the major means of manumission prevails, the slave population will accumulate to a considerable degree, and this is exactly what we find. In Korea where the rule applied prior to the Yi dynasty (though only partially during the Kory6 period), we find the most advanced slave system in the Orient and one of the most developed anywhere in the premodern world. Let us examine it more

closely.”! | |

From very early, birth became the major source of slaves in the Koryé period. Considerable attention was therefore paid to slave/free unions with a view to restricting any generational “leakage” from the slave population.

The rule of status inheritance was frequently debated and repeatedly changed, rarely in favor of the slave. The Chinese rule of deterior condicio

was rigidly applied throughout most of the Kory6 period and began to change only with the Mongol invasions of the thirteenth century. Amelioration reached its height during the reign of Kongmin Wang (1352-1374), and in the very last year of the Kory6 period the Roman pattern was made law. The year was hardly out when the new Yi dynasty reverted to the Chinese

pattern. This continued until 1669, when once again there was a change from Chinese to Roman type; but there was considerable unhappiness with the switch, and five years later the rule was changed back to the Chinese type. Between 1674 and 1731 the rule of status inheritance kept alternating between Chinese and Roman until the Roman pattern was finally enacted in 1731. By this time, of course, the landowning class had shifted from slavery to other more profitable forms of dependent labor. Exceptions were always made for the children of high officials and their favored slave concubines, many of whom were freed along with their children.

144 Slavery as an Institutional Process It is significant that in the two instances where the Chinese pattern was found among a primitive people—the To Lage and To’Onda’e tribes of the Toradja group of the Celebes, and the Merina of Madagascar—there was also the highest proportion of slaves, the servile group rising to over 50 percent of the population in some Toradja tribes.”

The basic Chinese pattern was followed by several of the barbarian groups that inherited the Roman Empire. In Visigothic Spain the principle of deterior condicio was rigidly applied.’ The same was true of Lombardic law right up to the end of slavery in the early modern period. In medieval Tuscany the legal situation was a strange mixture of Roman and Lombard law as well as local custom. The dominant pattern, however, was Lombardic, and the rule was that “children born of parents of unequal status took that of the parent of inferior rank: the child was considered a slave if either of his parents was unfree.””* Finally, there was the American case of South Carolina, where a statute of 1717 “for the Better Governing and Regulating of White Servants” con-

tained certain clauses that made the state’s pattern Chinese rather than Roman, as elsewhere in North America. All children of female slaves were to be slaves. If a free woman, white or otherwise, had a child by a slave, the child “was condemned to servitude for the ‘indiscretions’ of his parents.”””°

The Near Eastern Pattern The earliest recorded pattern of status inheritance, this was the dominant form among all the slaveholding societies of the ancient Near East. It was also the most liberal. The rule may be simply stated. When the parents were both free, the child inherited the father’s status. Where one of the parents was free and the other a slave, the child inherited the status of the free parent whatever the sex, as long as the father acknowledged paternity. In sharp contrast with the deterior condicio principle of the Chinese, the principle of melior condicio was operative here. As far as the marital status of the parents was concerned, this pattern required of free men and women only the power to acknowledge and pass their status on to their children. How they actually did so might vary: in the ancient Near Eastern prototype it was done either by marriage to the slave partner by the free person of either sex without any

penalty being suffered by the free partner or by freeing the concubine mother of the child, or by simple adoption of the child.°° Free women did not lose their freedom, nor were they punished for simply bearing children for slaves, although of course the practice was viewed with some disfavor and illegitimacy was strongly prohibited. A free woman could marry certain privileged categories of slaves and although, in view of her husband’s status, she could not become a mater familias, her children could not be claimed as slaves by her husband’s master.

Enslavement by Birth 145 The major exception to this generally liberal pattern of status inheritance was the Sirqu, the temple slaves of ancient Babylonia. This was a hereditary caste and could not be manumitted. In stark contrast to secular slaves, the basic Chinese pattern of inheritance applied rigidly.”’ The Near Eastern pattern is not only the earliest, but possibly the most widely distributed, pattern of status inheritance among the advanced premodern slaveholding societies. One of the main reasons is that the pattern was incorporated, in its essentials, into Islamic law and in this way was dif-

fused throughout the world with the spread of Islam. The Tuareg, as we have already seen, was one of the few Islamic slaveholding groups that did not conform to this rule. The key factor in the operation of this pattern among Muslims is the Islamic law of concubinage, which holds that a slave woman becomes free along with her children as soon as she bears a son for her master. Islamic law and practice were, of course, far less liberal when it came to the union of free women and slaves. Even so, it is noteworthy that the direct counterpart to the king’s slaves of ancient Babylonia, the Islamic Mamluks, could also marry free women; with respect to the ordinary slave

population, the prohibition in practice did not severely distort the pattern inasmuch as most slaves in most Islamic lands were females. (The exceptional Tuareg remain peculiar even in this respect, since among certain groups of them, where female status was high, free women could marry slaves if they chose.”*)

There were only two important European instances of the Near Eastern pattern, both of which emerged independently during the thirteenth century. In Sweden, when the laws of Ostergotland and Svealand were codified,

it was stated that “the child follows the better half”; as Joan Dyste Lind comments, “regardless of which parent was the free person, the child would be free.””°” It is still not clear whether this rule had always prevailed in these parts of Sweden or whether it was simply a part of the general amelioration of slavery that was taking place. In Spain at about the same time, the harsh

Visigothic law (which, as we have already indicated, was of the Chinese type) was changed to approach a pattern that was almost Near Eastern. The new rule was that the child of a Christian and of his Saracen slave or the slave of another master should be baptized, and if the Christian were the master of the slave the child also would be freed. Additionally, the child of a

Jewish-owned slave and a Christian would be baptized and immediately freed without any indemnity to the master.®” Religion obviously had a role

in this development, and not often from the best of motives. At the same time, we cannot rule out a possible Moorish influence. If religion was the sole factor operating, its purpose would have been achieved simply by insisting on baptism of the child of mixed parentage. The automatic manumission of the child of a slave concubine is so distinctly Islamic, and more generally Near Eastern, and so alien to the traditional pattern of Visigothic

146 Slavery as an Institutional Process and early Christian Spanish law, that one is not rash in assuming some influence.

In legal theory the majority of the slaveholding systems of modern Spanish America conformed to the Roman pattern of status inheritance, but in practice the rule was much closer to the Near Eastern pattern, with the important exceptions of nineteenth-century Cuba and sixteenth-century Mexico and parts of Brazil. The children of virtually all mixed unions inherited the status of the free parent (invariably the father). Where free women (usually freed women and Indians, but occasionally lower-class whites) cohabited with slaves, the women and their children remained free.” This may have been a continuation of the Near Eastern pattern found beginning in thirteenth-century Spain, but there need have been no historical continuity. A high incidence of miscegenation between masters and slaves, in combination with a less proletarian and more household use of slaves, will tend to generate a Near Eastern pattern. Even where slaves were used in a highly capitalistic manner and the legal norm was rigidly Roman, a Near Eastern pattern occasionally applied to the children of certain categories of slaves. This was the case, for example, in the British slave colonies of the Caribbean during the late eighteenth and early nineteenth centuries. Partly because of the shortage of white women in

the islands, it was normal for white overseers and even some managers of the absentee-owned plantations to take black and mulatto concubines. There was no legal requirement that these women and their children be freed and many, perhaps the majority, of them remained slaves. However, where ties of sentiment developed between the master and his concubine, the woman and their children were manumitted and even inherited from the father, so much so that frequent laws were passed in late eighteenth-century Jamaica limiting the amount of money that could be inherited by such children. While the group of so-called free coloreds that resulted from these unions was to constitute an important social class in the late slave and postemancipation periods, for the vast majority of slaves, including those who had temporary liaisons with whites and the resulting children, the Roman pattern prevailed. At no time was there any attempt to change this strongly

sanctioned rule.

The Sherbro Pattern Not many societies conformed to this type of status inheritance, where there was no fixed rule of status inheritance among the children of nonslave or “free” parents. Children chose the status of the parent whose line carried the greater status, although they tended to favor the father’s side. The same flexibility applied to slaves, except that it was the master of the slave’s parent who took advantage of the flexibility rather than the offspring. Among the

Enslavement by Birth 147 Sherbro of Sierra Leone “an individual’s claim to be a member may be made through either a male or a female link, although there 1s a patrilineal

preference. The matrilateral connection is usually stressed when the mother’s or father’s mother’s group is of high status.”°? The inheritance of slave status here is ambiguous. In keeping with the

dominant stress among the nonslave population, the child’s status was usually determined by the father. Women of ram (free) status who married slaves produced slave children, hence unions of this kind were unusual. Even though slaves were generally described as “those whom we would not marry,” a benevolent master wishing to reward a loyal and useful male slave would take advantage of the flexibility of the inheritance rules by providing

him with a wife from his (the master’s) own ram. The children of such unions could not inherit status patrilineally like most other slaves, but were considered “ram de, freeborn through their mother.” Although he is not explicit on the matter, we may deduce from Arthur Tuden’s discussion of slavery among the Ila that the Sherbro pattern held. It seems likely too that a Sherbro-type flexibility with similar patrilineal stress held for the tyeddo, or slaves of the Wolof of Senegambia.” Early Bermuda evinced an illiberal version of the Sherbro pattern that applied only in cases where a child was born of slaves belonging to different masters. The first such child went to the master of the mother, the second to the master of the father, and so on. Only quasi-Sherbro, this example conforms to the pattern in its flexibility but differs in that the parent’s gender determined only the master of the child, not his or her slave/free status.°’ The practice was highly reminiscent of the distribution of children of serfs belonging to different masters in medieval France.** One other variant of

the Sherbro pattern did influence the status of the child, and it was the practice that prevailed among the Iban tribe of Borneo. From Brooke Low’s

classic account of the group during the nineteenth century we learn that “where the parent is free on one side, and the other parent either an in or outdoor slave, the first child follows the fortunes of the father, the second that of the mother, and so on in succession, and this rule is unalterable.’©

|i tThe ®® Acquisition |

SLAVEHOLDERS ACQUIRED SLAVES either directly, using

any of the means enumerated in Chapters 4 and 5, or indirectly—that is, from third parties via trade, gifts, or closely related payments in kind, or when the slaves were used as money. The most important of the indirect means of acquisition are considered in this chapter.

External Trade External trade always played a major role in the indirect acquisition of slaves. Few would challenge this statement in the case of the advanced premodern and modern slave systems. What may seem surprising, however, is that it holds too for the most primitive of the societies where slavery was important. Slaves often constituted the earliest article of trade, especially of external

and long-distance trade, among primitive peoples. The only commodity simple peoples could usually offer to more advanced peoples for the luxury goods they desired was fellow human beings. This becomes evident in stud-

ies of the indigenous West African markets and trade. Summarizing the findings of his own work and that of his colleagues, Claude Meillassoux concludes that the slave was both a commodity and a producer in West 148

The Acquisition of Slaves 149 Africa. Sometimes the slave was involved with trade purely as a commodity,

especially in the destructive trade with the Europeans. In intra-African trade, the slave figured both as commodity and as producer. Slaves were also

vital in long-distance overland commerce as porters and in the capture of more slaves.' This, however, was true mainly of the more advanced societies. In the simpler lineage-based communities “goods circulate through a network of kinship, affinity, and clientage, through prestation, redistribution, or gift exchange. Wealth as an instrument of social control is a privilege of rank or of birth.”’* In such small-scale communities trade was sometimes

a threat to the established order and was therefore circumscribed. Hence imported goods “acquire a social and political content which makes it difficult to transform them into trade commodities.” Slaves were prestige goods and at best a “means of social reproduction,” rarely a means of production. In the Sudan the use of slaves as producers for external trade was always accompanied by the rise of both a warrior and a merchant class.’ The archeological data on neolithic Europe strongly suggest that slaves were among the earliest articles of trade.* One of the most striking articles

recovered from the Llyn Cerrig Bach find in Anglesey was a slave-gang chain, leaving no doubt that slave trading was a well-established practice in

the La Tene culture of the Celts during the first century of our era.” This trading continued down to the Viking period, as we shall see. Among the Indians of the northwest coast of America there was a welldeveloped pattern of long-distance trade in which slaves were the principal items of commerce. One reason was the propensity of slaves from neighbor-

ing or even moderately distant tribes to run away and rejoin their own groups. Thus the price of a slave was largely determined by the distance of his point of origin from the final buyer’s home. Large-scale slave marts dotted the coastline. The Dalles, a slave mart, became the Delos of aboriginal North America. Trade moved in two directions: from the south up to the Dalles came slaves who were exchanged for other goods that came from the north; the slaves were then traded farther north.° It should by now be clear that slavery was intricately tied up with the origins of trade itself, especially long-distance trade, the bartering of slaves for prestige goods often being the sole form of commercial activity. As the demand for slaves grew, slave-trading systems expanded in both organizational complexity and distance between areas of recruitment and areas of use.

Throughout recorded history, even to the first half of this century, slave-

trading systems have always existed to meet the widespread demand for slaves. Five systems stand out in terms of the volume of trade and the distances involved: the Indian Ocean; the Black Sea and Mediterranean; the medieval European; the trans-Saharan; and the transatlantic. A few general remarks on these trading systems are warranted.

150 Slavery as an Institutional Process THE INDIAN OCEAN TRADE

Perhaps the oldest slave trade of all, the commerce in the Indian Ocean had both an east-west and a north-south axis (see Map 1).’ As early as the Eighteenth Dynasty (1580 B.c.) ships were sailing from Egypt to northern Somaliland (the land of Punt) with the specific aim of obtaining slaves.* There is reference to trade in slaves from East Africa to Alexandria in the early second century A.D.’ The large number of black slaves in the Persian Gulf area attests to a slave trade preceding the rise of Islam. This trade was intensified as Islam grew, and Arab traders for slaves in increased numbers began establishing posts on the East African coast as far south as Zanzibar, probably as early as the ninth century. The volume of trade, especially during the nineteenth century, was much larger than is normally believed. Indeed, the East African trade during the nineteenth century was significantly larger than the Atlantic slave trade during the nineteenth (or seventeenth) century. R. W. Beachey estimates a total volume of 2.1 million slaves exported at that time, not counting the “fringe numbers.”'° Between 1800 and the mid-1820s approximately five thousand slaves per annum were exported, and about the same number held during the last quarter of the century. During the middle half of the century, when the trade was at its peak, some twenty thousand slaves each year were exported south of the Horn; and between the mid-fifties and the end of the seventies more than thirty-five thousand were exported annually from the northern half of East Africa. The vast majority went to the slaveholding societies of the Middle and Near East. For example, for most of the nineteenth century between fifteen and twenty thousand slaves were exported annually into Mecca and Medina from the African part of the Red Sea. The Portuguese may have taken as many as two hundred thousand slaves from the region during the last decades of the eighteenth century, and approximately the same numbers were absorbed by the advanced plantation systems of the Arabs on the East Coast itself. Other authorities estimate the total volume of the trade between the years 800 and 1800 at three million. Thus during its entire history the trade involved the acquisition and sale of approximately five million persons.

THE BLACK SEA AND MEDITERRANEAN TRADE

The Black Sea and Mediterranean slave trade was another of the oldest and most important in the history of slave acquisition (see Map 2). Despite its importance for the ancient economies, we know very little about it other than that it rose to prominence toward the end of the seventh century B.c. Before our era the southern regions of the Black Sea and Asia constituted the single most important source of slaves, although significant numbers also

Pe —

. 0_

Basra

AYoS,cain Nexandha NY

" a Zubaic ° M a 1 nnn ~ \ a \ Fj 8

ia Se = SEES. , wv Ee‘ee —o - BE . ~ ADE hE an L G Magdalatolal‘ =yg ae—>Sys. 4MEAS . 7" BOSS ~~ =ELIE] 17. : 883 LAL su bt Cc n ° oe .? 3ri |5oz w CeRY ws Wee\Ssseqp \% ol MLEG SH2leg AN ee ASSISTS] -. iLS Y, SSE CM SRP SA AeISASE eo . 7 Sh wT. aWO :ve PES NGO ’ :7oeWr sv 3 FE EO VIS TS ee 4 NT x Soc = aA, She gr e az oft} . 7 Os: oa Ee 4.2 atte Ans SN i” eo? ° e 8 ty. . . re ed

a 5° rs Ee lt SRP Ds re? ee oe : Be eS! BL ee se cx. 2 Sf = & Zz CI BFS. Lg ® : soe} gS sé s 2 ~~ ¢& je > oe POET ry wi — oe ‘ QU ; . wet & fae} ‘“ss, ee tl . ° +oO. }= aa

bi} aeo ucek &ria °aN SS 3(yr we OL 2 ee |SIC,&; ~~2SS©& TON Cha £ é | ey 9 O os Ro RE EN. Saori, Bz tt MARES SS ENV NOSES or Rae 24 Tre Var < oC od ASS): 4 Me = hal~~ rg Za ~OIW" , os‘iets e a4 ‘ thpe, - ‘S* ereS ° .. an Fe rR

FS pel i_ aee | BR x AzESE ayeRf«‘©2SPsleaghew * es a a* say SsD Se oO esrUfSNVCAN & “KhSee, FSR -V«a UO

fH N\A Eee ; SA | aS ey a wee Redo oS | wU< Te a oe Bs AHESS 2 Ll ee] BB 8 £5

NY, =-- Qa ~ Ry . ~~ f° , 2 v— e ° vo & . \» . . °. aan ate ° . e . .

‘= o or Nt SN NY EUG iss © ds 2 CASTRS Wa fi °° , 6 . ‘ . or © Wy

: on \ 25 ig ' ye ei 4 bait KN |BA ws St ~NBR . . . . a —_= oo} QO. A i Noes 9 Sal gg Xe ~ (00 re g fOr . Y * =\S . ar: ro) —_~ re\ 8 MES OF NS SPY Wake (cre 1. Bl ose A OX2\ Pe = SCOR UST TES &SRE ec “ki s\CANi© 'YH How not mW Se ZR OM ge eg ee || BES ater ws AN=Q2x \e—~ SS a ZRa eeif. ee2ten cechs oe\o\= Ye GOUT bd uw ee o> (4OD ~ TDD ne) Ss « yo \ < S ete GEES +TesoRR mA Ps t 6 © ZX ive ‘Me $= : : ‘ . 2 ™ ™| : . “wd ye fe ss S 4. Ei Sug ee Se RNs Dbor pb oS Se S

®A. i ndG& _— *“\> ~them & % ook fp .©-..£bd % VU Oo me! ;,\~lg rN,Ca 2 }Mrs WM soe ng —_ ' oe TetiN SN {LW : >&yUL :6L — eeny a, ~r; \,“Deere (2 1pat) s Ns,» 3>,FB) wa~~£ 6

ane: neal ae 82 HG (g', fe WEE eeages oS x ~cheat u Zz 7enASY ' 4 ha ly >[geee ~w >& ~~A*,GM: ‘ Ys fer So 5&& Q SER Py aLe “hoy 2 a Yo te (SSse © NCCT 5 g +H Se oe meee, As (S Sass] \SG = Ssaem, 5, £ ..§ = wale @ e}&LY Q 6“Ih,i=EO ro)Bet] agi 7A ug *e, | ENS OATS ‘.¢ SEE. a Hees | RSENS UMN LSE MS 4oa SPL28e5: "+ && =. —Mee ™:The aeLa RS~% Weees A.S1Ko ° @) oT

: e *;EL iy=fi \i O9Sg4 aonSX Ae \2 ~ pF He & Pa ae] ie)=& = SeHlser, df jslsyp QTE FFCK. ~Qg V O° CEVENRS —~ SISA Kz] [ek* SS SBba’ BAR a=

: . eee See eee > ian ' » >

154 Slavery as an Institutional Process | the African middlemen on the west coast of Africa eight hundred years later, were to be both victims and dealers in this trade. They raided and traded in the interior for slaves, then sold them to the Viking seaborne merchant pirates. The port cities of Cardiff and Swansea may have begun as slave-trading marts. But the two major centers were Dublin and Bristol. From these

depots and from their own home bases, especially Hedeby, the Scandinavian merchants spread out all over Western Europe, and frequently as far as the Mediterranean, selling their human cargoes. The Vikings did not scruple to raid their fellow Scandinavians. A considerable number of thralls were taken from neighboring Nordic peoples. The

Icelandic poet Valgard describes a raid on the Danes by a mixed group of Norwegians, Danes, and Swedes wherein “the Danes, those who still lived,

fled away, but fair women were taken. Locked fetters held the women’s bodies. Many women passed before you [the conquering king of the pirate band] to the ships, fetters bit greedily the bright-fleshed ones.”'’ Nor did the Vikings hesitate to sell Scandinavian as well as Slavic and Celtic slaves to the Muslims. On the western route some of these slaves were taken southward to Lyons and on to Spain, where many were again traded by Muslim and Jewish merchants farther south and east to the Muslim states. There is also evidence that there was some movement of slaves from south to north, for the “blue men” who appeared in Ireland in A.D. 859 were almost certainly African slaves brought there by the intrepid Vikings from Arabia or some other part of the Muslim world.'® After the middle of the ninth century the heaviest traffic in slaves was on the eastern route. It was then that the Volga and Dnieper rivers were opened up mainly “as slave routes to the eastern market.’ Birka, situated on the northwest part of the island of Byjork6 in Lake Malar, not only became the pivotal point in this trade but until it was abandoned in about the year 1000 was crucial for all northern and central European trade.”° From here slaves

and fur could be sent south to Gotland, the South Baltic, and Hedeby, which dominated the western route; silver and Slavic slaves could be distributed north and west for use in Scandinavia itself; and, most important, fur and a few Scandinavian slaves could be sent through the Gulf of Finland to the Volga, then south through Russia to Bulghar, where they met Muslim traders from the east. It must be emphasized that most of the slaves traded south did not come from Scandinavia, or even from the Slavs of the Baltic, but were gathered on the way down in terrifying raids on the native Slavic populations that lived closer to the market. The distinction between warfare,

piracy, and trade was completely blurred among the Vikings. As P. H. Sawyer points out: “Most if not all of the Kuffic silver in Scandinavia was acquired by way of trade. The goods that were sold were probably gathered by violence, and the silver that reached the Baltic was at least partly distributed by piracy, but there was also commerce.”””' It was the Viking raiding

.‘: : |. :"@ o) oo “oe,y's QO - e., °ee =x =, oma : Shel fs S | = : ori x==“>22. Py.,ne af 2 + aay 3 :Wee . Linetyoy3a=@ | 25 red / aI + mt ~- . oe

oc : ff \ @ S 7 ¢ —~ > i1|LP>AReoFPreexcssten” 53Se=se‘°.we: 3 +o ~ 44,'moe “sy RAF osees “ 4 SXS YW) v7 x we . S2R fo SSK Sos a=

Oya besele 8 LFS ZL [3 ” » = S A ; al y & A } ye cA fe“A ‘Neh w“ & ~S cnix8re) LA a3. . Ly :“oy SQN WS =we 7reoo} , * * CJ ’ vet ee Ss 4 oe, ~ BRS FS 5S roe XN \fD Pa G b ro} y . ye a 4) ‘c xa. {| NN AS sSs24itSO Oo aS * ° >, . ear N=. By ~8 > *. i we c SSS ¥EN ta ret eon < .. Lg Bees SS %& At Rr UOT. ie woge ey Gs oH aa - ‘ BH © g S ass @ = Se 7 ’ ae oh fo uw m .yrs~ fr: :vwe - —_Dad “ r 4“ffNeWA. |lps aieSit Ww Z nese iS x:“\.:ba a| | nae yx ss ky, 36S x ea ; Fqo 3pita 3 1Y ee S38 Hi Ssasd BS 4wy em=oS ‘@) (a | ie 3 7 re) 5 5, ee | ~~ |||j =4S x eX ¢ S “Soa 7 2 fof? | wy ON O >Ko ashe . F=BR wl So $. 23 —:tepS ae >-.ote Fa af-—= Se =. 4 5 c woe, Ne fi “td iitm EES eS PS bY’— ~FS, >; .oy of! mn x" wee De SS Cc esi aa . ; me i. seat AS 2S pte \ 2 ee: i‘SG S92=a weir Y. aa Me ZeCg SN~ ik sews ie\Ea8 WI Of \PY eR

i.®t. oot > f \ SS © !| Ne Se,:MF °SYPHRE sy fis # ~. . = i> EST Og ese —\ => . FSS ena ©. ails-_ Sineoe"2 «6 7 2=Soy ‘a Qa 2 ‘ “eee ~ Cc are ’ i“ ® -R: 8 . ~~ on = = x aed x “"S . 4 oy “ 4 ey Ch Wes>ye ME“3 LSSe ~ :; Ss ee B2EeS FQ APS SSP

Og iD2,:|sakSe=_¢sea\=stv“rea} 2. 234iS $0 | ro By — = o | . 4 o= a; . ern SG Ags 4 a . ‘aBrasy y .. = ~ oe Oo oi a s Ss —_ = 5ees SE s Sy Up e— CSS ais * = “ Dy ODa . dar 4 35s 4 aeosx.‘ ea . 4 ;= ow = Cc 4 oad 3 He 2G ? ~ 2 “Ts v ar w= : fay toy ta ~ f— la w $ q o 26 Ge ole : G9 S ” wo Siz : oO 3 0 = a. 8 aN x wo oD -oN ~~ x, @ De . = ke cob) © oO ) ™? 2s . ~ AL ® QD wa = A= ” ci >x ns 43

: ) aa Je Le z oO

Bee eS &. S

156 Slavery as an Institutional Process

and trading of Slavs that led to the common root for the term “slave” throughout the European languages, not to mention its use as one word for slave in Arabic. Still, the Slavs like the Celts to the west paid the Norsemen back in kind. Their raids became more daring, especially after the decline of Viking power. In 1135, for example, Slavic raiders plundered the west-coast Swedish city of Kungahalla and took some seven thousand captives, all of whom were sold into slavery (although, as with so many other contemporary figures, this number seems greatly exaggerated).”* It is impossible to make any estimate of the volume of the European slave trade. Recent studies have tended to underplay the destructiveness of the Viking raids.” It is now commonly agreed, for example, that most of the raids into England were preliminary stages of settlement. It seems likely too that most of the slaves captured by the Scandinavians in their Western European raids were used by the raiders themselves, either in their home societies or as laborers in the regions they settled.** This is consistent with the best explanation for one of the Scandinavian numismatic mysteries, “the rarity of English and Frankish coins in the ninth century,” which Sawyer finds “extraordinary.”*? His reasoning is that the English and Frankian coins gathered in plunder and as tribute were used as capital for settlement in these same areas, so that the coins never reached Scandinavia. In keeping with the Viking pattern of mixed raiding, slaving, plunder, and trade, a great many of these coins were also used to buy slaves to assist in settlement of the plundered areas. This is reinforced by recent evidence on the carrying capacity of ninthand tenth-century Viking ships.*° The extravagant estimates of earlier scholars have been considerably scaled down: for long raiding voyages the aver-

age ship during the ninth century carried no more than thirty-two men. Ships in Scandinavian and other Western European home waters carried more, but not a great many more. What we can conclude from all this is that in Western Europe the heavlest traffic in slaves involved short distances. Most slaves could have been captured and bought for the Scandinavian home market along the Scandinavian shores. Similarly, the slaves of the settlements in the British Isles could have come mainly from other parts of this area, although there might also have been a shunting system, essential to avoid excessive running away, similar to that of the indigenous slave-trading systems of West Africa and among the Indians of the northwestern coast of America. The “blue men” of Ireland and the “bright-fleshed,” flaxen-haired boys and girls in the homosexual and heterosexual harems of the Muslim East must have constituted an insignificant fraction of the total volume of the European slave trade. It is, as noted earlier, impossible to offer any sound estimate of the volume of the European slave trade during this period. The following guesses are at least suggestive. J. C. Russell has estimated the population of Western

The Acquisition of Slaves = 157 Europe around 950 at 22.6 million.*” My own estimate is that the lower limit of the slave population of Western Europe at this time was [5 percent of the

total. This is highly conservative; in England a little over a century and a

quarter later, when slavery was on the decline, the slave population averaged almost 10 percent of the total and in the western regions over 20 percent.” The slave population in parts of Scandinavia was likely to have been at its highest point during this period. The Icelandic settlement had reached its climax by about 930, creating a critical labor problem in both Iceland and Norway. In other parts of Western Europe the population decline during the first half of the tenth century would have intensified the demand for slaves.*” Thus we can estimate the total slave population of Western Europe at this time at approximately 3.39 million. From what we know of slavery in the kind of societies that existed then in Western Europe, we can confidently assume that the slave population was both socially and biologically reproductive. The annual demand for slaves, in that case, would have existed only to meet the needs of new settlements and to compensate for social “leakage” due to manumission. This could not have exceeded | percent of the total slave population in any given year.”’ Hence the annual volume of the Western European trade could not have been greater than a total of 33,900 slaves at the middle of the tenth century. It is difficult to even guess at the volume of trade on the eastern routes, although it certainly must have been much greater. At the very least, then, the total volume of the trade at its height must have been between 67 and 68 thousand slaves annually. THE TRANS-SAHARAN TRADE

The trans-Saharan slave trade lasted for almost thirteen centuries. As Philip D. Curtin has pointed out: “Islam and commerce was first associated in West Africa because Islam came across the Sahara carried by merchants, and contacts on the Sahel were between merchants.””*' From as early as the ninth century, highly profitable trade diasporas were established to take advantage of the demand in the North African and Mediterranean states for African slave laborers and African goods. The four main routes to the Mediterranean coast (see Map 4) were from “Timbuktu to Morocco, Kano to Air and Ghadames, Bornu to Fezzan, and

Wadai to Benghazi.”*’ The slaves carried in this trade eventually found themselves in almost all the Islamic slaveholding societies of North Africa and the Middle and Near East, with the main areas of ultimate purchase being Egypt, Morocco, Libya, Tunisia, and Algeria, in that order. Estimates of the volume of the trade vary considerably. Curtin claims that it was particularly large during the last quarter of the seventeenth and first half of the eighteenth centuries,’ while most others view the nineteenth century as the period of greatest traffic. The most systematic evaluation of the direct and

e ° o Oo .

oS|fi aas zt ba >-x.otY a = >an i j7 ; | o a

Z¢

> Oo

ge = r | |g j--e¢ \ x gi {|}@,; bs a| i4on iPn: // ava {4xWw ~~ “ x tig

f) =f ran ' (tl 3| — “7 ~~@ x || zt , \| ro) so\ ~ Le y \ a‘\ ‘“ | ob wen

Yw| ~. x riow wy) z (\/ it\>“~ 7? | 1 -~~--—-~< © 26 is \ a \\ { s&s /(1yj Vo oc K Ss a 5 wu \ ‘\ . 2 oS ra i 2 IN \ a ‘\ { , 2 an oS 3/ s . \ 1 \ @ CO _%e ns \ le nn re a~ a) a « (9) Lis Nf e Xo 6 oO 4 “ 2 < Ww) ACH a ia = \ = ” im ZB J Vv} / \ = ba Oo 2 gf oy rt 4 g | 2 Sao be = Yk Zql ™ : > a \ a \< ‘ ” aT} ~x ne SY{ @) ™ 6NI™ Zt< °KS) 5ss is ag :

zt ~ PO oO _) , «“ @aN i— ‘tT“—* Or3| !1j/ /2ag ©6 a5 “Mw

—: aeAEaé * | la

=. Loe

\ \ re” , I / oS zOo \ 4 / N > \ Vt XxX & .. Oo. @ \ «a ef \ ek O-N GS 7 q x + 8 go Ete aud

~~. 3 Z, 5 \D N me ID rT _ OQ wy rr 4, © . t/a , | > a ©af 8 5-7.

8m x 7 , a ed a sor 5 x 4 o ao a, S ’ tg 7 or ae z= sé a 2¢ ~ xf \ ro) S yo 8 38 ra r/ > Of \ = co a ji ‘ne o| e¢ a0 se. \ ° - ° aes e \< / ~~. uu 12 \

> 4 ano ’ ? 3 é ) ee oo a | EH & © 5 mH

~

.S

a

The Acquisition of Slaves = 159

indirect statistical data is that of Ralph A. Austen, who estimates that between A.D. 650 and 900 some 450,000 persons were transported in this trade; between 900 and 1400 approximately 2.8 million persons; another 2.4 million were carried away between 1400 and 1800; and during the nineteenth century about 1.2 million.” It is striking that the total number of persons acquired in this trade is well over half those taken to the Americas in the At-

lantic slave trade, and that when the East African total of five million per-

sons is added, the total volume of African slaves acquired by Muslim masters is greater than the total acquired by Europeans in the Americas (even after making allowances for upward revision of Curtin’s estimate of the transatlantic trade). There were, however, important differences. The rate of acquisition was much slower in the trans-Saharan and East African trading systems; and the overall mortality rate was higher. However, the pattern of mortality also differed. The mortality rate en route in the Saharan trade was between 3 and 7 percentage points greater than that of the Atlantic trade, but the mortality rate before embarkation on the middle (transatlantic) passage was much greater than that involved in the enslavement of persons destined for the

trans-Saharan passage. The proportion of men enslaved was also much larger in the Atlantic trade. And, not least of all, the experiences of Africans after reaching their final destinations were radically different in the two systems: in the Americas the slaves were mainly absorbed in capitalistic systems as a rural proletariat; in the Muslim world they were used largely as

domestics (although one should be careful not to underestimate the nondomestic and rural uses of slaves in these societies). The trade declined toward the end of the nineteenth century, under direct pressure on the suppliers by the European colonial powers in Africa and diplomatic pressure on the Muslim states of the Mediterranean and Turkey. However, the slave trade was never completely abolished. A UNESCO report on Mauritania in 1960 claimed that a quarter of one of the largest tribes of the country were slaves, many recruited through the trans-Saharan slave

trade.” A much reduced but significant flow of slaves also found its way across the Sahara to the states of the Arabian peninsula, which continued to hold slaves right up to the 1960s and may, indeed, still have a few. In 1960 the trade to Saudi Arabia ran “from villages in the French Sudan, the High Volta, the Niger Provinces and the region of Timbuktu ... across Africa to the coast at Port Sudan or Suakin and thence across the Red Sea by dhow to Lith, a port south of Djedda.”*° THE TRANSATLANTIC TRADE

The last and greatest of all slave trading systems—the Atlantic—began as a diversion from the trans-Saharan and Mediterranean systems. The earliest

160 Slavery as an Institutional Process groups of Africans to land in the New World came from the Iberian peninsula, to which they had been delivered by the traders of the Mediterranean system. And the earliest group of Africans coming directly from Africa was recruited from the Senegambian coast, from traders primarily involved with the trans-Saharan trade.*’ Very early, however, the demand for slaves in the New World outgrew the capacity of these two ancient slave-trading systems. Almost all the Western European peoples were involved at one time or another with the enormously profitable transatlantic slave trade. The Scandinavian role, though not insignificant, was minor compared with the major slave-trading states: the Portuguese, Dutch, English, and French.** Although Spain was an important consumer of the slaves carried by this traffic, its role in the actual commerce was small. This was not because of humanitarian considerations, but simply because its resources were stretched too thinly controlling its vast empire in the Americas. The Portuguese were the first to develop the trade on a significant scale, but by the end of the sixteenth century their monopoly was being seriously challenged, mainly by the Dutch; by the end of the seventeenth century the English and French too

were heavily involved. OS

The slaves were recruited almost entirely from the western coast of Africa, from the Senegambian region down to Angola (see Map 5). Except during the last decades of the trade in the nineteenth century, the coastal belt of tribes (extending no farther than two hundred miles inland) was the principal source. Certain African tribes therefore lost far more people than others. Up to the end of the eighteenth century most slaves came from the tribes of the Guinea coast region, an area that, in spite of its large number of tribes and languages, had marked underlying cultural uniformities. In the next century most slaves came from southwestern Africa and, to a lesser extent, from Mozambique and central Africa. _ A whole new subspecialty of historical studies has developed around the problems posed by the demography of the slave trade, largely stimulated by Philip Curtin’s census.*’ Curtin’s estimate of 9.5 million persons (plus or minus 20 percent) imported to the New World by this traffic is widely used, but most recent studies relying on archival data have tended to increase his figure. It is safer to say that between 11 and 12 million Africans (plus or minus 20 percent) were imported to the New World. _ Figures 6.1 and 6.2 summarize the most striking features of this traffic. Most slaves came to the New World during the eighteenth century. The United States, which imported the smallest percentage of Africans, by 1825 had the largest proportion of slaves in the hemisphere. The Caribbean slave societies, on the other hand, which imported over 40 percent of all Africans brought to the New World, had less than 20 percent of the slave population of the hemisphere by 1825. The difference is due to the remarkable rate of

natural increase of the U.S. slave population compared to the equally

Pere eScece Ee OSes .SES See — Bes ee Bae eee b @ CS / Coe Sees ON Bon AS rt 5] m po ae Sees SeeGs Ss ~— n USE oe “S —_ OQ i" fro. oe ae ce Pe carn & LY ron See os SRE eee ens ee;aty ne ons og *sail sg Sonny wy EEE |bNene’ = . g-~ bx a=ps —— eee m YQ wy BeeeeEee = Jc6 86— eS rae)cad, < Ce

ed le 3- of=

ES eae sh SoSe SeesPine BES ee _ Ospee J See Berean

aSecsae Bore SEES oo ce stn Sere Bey Sa SoBe Sees ee eee toes 585 See pieiecatie set sae eee eee _ oy

Sa CC 3 ae re Recor eeecoe EON Sd _ oe aaa ee Pee SEE a o S Rees sna Se see — a eS sk: Sees CREE eee ee Sees SSE S

162 Slavery as an Institutional Process

1451-1600 : 275,000] p Old World New World Total imports 11,544,400 Figure 6.1 Imports of Negro slaves by time and region. SOURCE: Adapted from Robert W. Fogel and Stanley L. Engerman, Time on the Cross (Boston: Little, Brown, 1974), vol. 1, p. 16. Revised estimates are based on Roger Anstey, “The Volume and Profitability of the British Slave Trade, 1761-1807,” in Stanley L. Engerman and Eugene D. Genovese, eds., Race and Slavery in the Western Hemisphere, (Princeton, N.J.: Princeton University Press, 1974), pp. 3-31; Patrick Manning, “The Slave Trade in the Bight of Benin, 1640-1890,” in Henry A. Gemery and Jan S. Hogendorn, eds. The Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade (New York: Academic Press, 1979), pp. 107-141; D. Eltis, “The Direction and Fluctuation of the Transatlantic Slave Trade, 1821-1843,” in Gemery and Hogendorn, The Uncommon Market, pp. 273-298; and Herbert S. Klein, The Middle Passage: Comparative Studies of the Atlantic Slave Trade (Princeton, N.J.: Princeton University Press, 1978).

remarkable rate of natural decrease of the Caribbean and other New World slave populations. These shifts in the proportions resulted from both higher death rates and lower birth rates in the Caribbean, although the latter was the more important. Better diet, shelter, and general material conditions account for the higher U.S. birth rate. Continued reliance on the slave trade — partially explains the higher death rate in the Caribbean, since the larger proportion of Africans compared to Creoles meant a smaller number of women (less than 40 percent of the Africans brought over were women) and an older population. Differences in lactation practices between the two regions also account for the higher fertility rate of the American slaves; the greater fertility was also in part the result of a lower relative reliance on the slave trade in the United States.” Regarding the organization of the slave trade itself and the experience of | slaves on the middle passage, recent works suggest the following. First, there

The Acquisition of Slaves 163

40 ; $30 .

50 Percentage of Slave Imports, 1500-1825

é Ee

|| Percentage of Negroes in Western Hemisphere, 1825

Ned ooaoes 20 — ] 0S

0 Eee _ Ze = Bes |

United British French Spanish Brazil Dutch, Danish, States Caribbean Caribbean America and Swedish Caribbean

Figure 6.2 A comparison of the distribution of the Negro population (slave and free) in 1825 with the distribution of slave imports, 1500-1825. SOURCE: Robert W. Fogel and Stanley L. Engerman, Time on the Cross (Boston: Little, Brown, 1974), vol. 1, p. 28 (© 1974 by Little, Brown and Co., Inc.). By permission of Little, Brown and Co.

was a remarkable similarity in the organization and functioning of the trade among all the European nations who participated in it. The most individualistic trading nation was Portugal, which relied less on the triangular pattern that linked Europe, Africa, and the New World into a single trading system, and more on a direct two-way traffic between southwestern Africa and Bra-

zil from the second third of the nineteenth century on. Second, mortality rates were equally high among all the trading nations, one estimate of the average annual mortality rate during the eighteenth century being 13 percent. However, there was a general decline of this rate during the eighteenth century. An important recent finding is that “tight packing” of the slaves in the ships was not the major cause of mortality on board, but the length of time at sea, the quality of food and water during the passage, and epidemics and health conditions at the point of embarkation in Africa.*' (This does not take account of mortality in Africa before embarkation, a point we shall return to later.) The price of slaves in Africa averaged between 3 and 4 pounds sterling

164 Slavery as an Institutional Process during the last half of the seventeenth century, rose during the eighteenth century to a peak of 18 pounds in 1740, declined sharply during the 1740s, then oscillated upward to about 17 pounds in 1770, when it began to fall again.** Richard Nelson Bean, who has done the most thorough work on these prices, has drawn several rather controversial conclusions from them. One is that there was a direct relation between price fluctuations and the varying volume of the trade. Wars, changing transport costs, and other political factors were the shifting forces that influenced both supply and demand for slaves: “In each case the market acted as would be expected from the basic maximization postulates of price theory. This was as true in the response of African slavers to the stimuli of changed market conditions as it was in the reactions of the British planters.”*° Bean’s second major conclusion is derived from the first. He argues that if the slave trade had created greater population losses in Africa than the rate of natural increase, there should have been a much larger rise in the price of slaves coupled with an equally sharp decline in exports. That, he claims, was contradicted by the joint increase in both price and volume. Not everyone would wish to so confidently draw this conclusion from price series data. Even more controversial is Bean’s claim that the “European contact, even inclusive of the drain imposed by slave exports, may well have meant

that there were more Africans alive in Africa in 1800 or 1850 than would have been the case had all Europeans left Africa strictly alone.”** Curtin has pointed out that in Africa slaves were sold by their African middlemen captors “for much less than the cost of reproduction,” and the appropriate economic model for estimating the effects of the trade is not the “fishery industry” preferred by Bean but “burglary.””’ Other later studies on the vexatious question of the effects of the trade on Africa strongly contradict Bean’s ex-

traordinary claim. Henry A. Gemery and Jan S. Hogendorn, in particular, have shown that even with the very best set of assumptions, West Africa suffered not only severe net economic losses but incalculable demographic and social losses. Millions of slaves, for instance, died between being captured and being forced on board the slave ships. In order that the slave masters of the Americas might acquire 11 to 12 million slaves, at least 24 million persons were originally enslaved in Africa.”

Internal Trade Apart from the external slave trade, masters could acquire slaves by means of an internal slave trade. This type of trade, drawing on locally born slaves, was actually quite unusual in the history of slavery. In premodern slave systems, however advanced, there was a universal reluctance to sell locally born slaves. Such a sale was usually a form of punishment in most slaveholding societies, especially in the case of persistent runaways. In imperial

The Acquisition of Slaves 165

Rome, for example, some slaves were undoubtedly acquired in internal

trading but the numbers involved were insignificant.*’ , It is only in the Americas that one finds internal slave trading on a signif-

icant scale. The two most notable cases are the United States South and Brazil during the nineteenth century, especially after the conclusion of the Atlantic slave trade. Recent studies indicate, however, that there was a minor internal slave trade in Jamaica after 1807. With the expansion of cotton production and the corresponding shift of economic focus within the U.S. slave South from the northeastern to the central and southwestern states, an interregional movement of slaves took place. It has been estimated by R. W. Fogel and S. L. Engerman that between 1790 and 1860 some 855,000 slaves moved from the northeastern states of the Old South to the new Deep South. The same authors, somewhat more controversially, estimated that only 16 percent of these slaves were sold from one master to another, the remaining 84 percent actually moving with their masters.** The popular view that slaves were specially bred for this trade has been vehemently opposed, although it still has supporters.” From very early on there was an internal slave trade in Brazil. It was of minor significance until 1850, when the abolition of the Atlantic slave trade combined with the rise of the coffee plantation stimulated an internal trade that according to Robert Conrad was “strikingly similar to that which developed in the United States under comparable circumstances.”°” The superficial resemblances were indeed remarkable. In the same way that the rise of cotton created a demand for slaves in the Deep South, so did the rise of coffee create a tremendous demand for slaves in the central and southwestern states of Rio de Janeiro, Sao Paulo, and Minas Gerais. In like manner slaves were shifted from the older northeastern states to the newly opened ones. Conrad goes so far as to argue that the drainage of slaves from the northeast of Brazil “compelled” farmers of this region “to make an earlier transition to a free-labor system” and encouraged them to challenge the whole system of slavery.”'

This thesis, however, has been challenged by Herbert S. Klein, who argues that the interprovincial seaborne trade in Brazil transported only five to six thousand slaves annually during the 1850s, 1860s, and early 1870s;

that the main function of this trade was to supply a limited number of skilled slaves from the northeast and extreme south to the south central states; and that this trade was simply not of sufficient volume to account for

either the decline of slavery in the northeast or its increase in the central states.” It is interesting to note that Klein’s revisionist thesis is similar to that of Fogel and Engerman with respect to the American South. The two slave societies are still held by Klein to be similar in the pattern of internal slave trading, but in ways quite different from the traditional positions of Conrad

166 Slavery as an Institutional Process and Frederick Bancroft.’* The cliometricians have deemphasized the interregional movement of slaves; yet it is important to note that they have not

proven (or attempted to prove) that there was not an important internal slave trade. Rather, it would seem that the internal slave trade was overwhelmingly intraprovincial or intrastate rather than interprovincial or interstate. Unfortunately, interest has centered more on the problem of interre-

gional movements than on internal slave trade regardless of the kinds of movement involved. Available work on both Brazil and the United States suggests that when intraregional trade is taken into account internal trading was indeed significant. The scale of the internal slave trade was much smaller in the Caribbean, although the pattern was similar. The end of the British Atlantic slave trade in 1807 was followed by a small, insignificant interisland trade.”* In the case of Jamaica, the internal trade was not a great deal more significant than the interisland trade. Between 1829 and 1932 only 4,838 slaves moved from one parish to another, amounting to only 1.5 percent of the total slave population. As in the United States there is no evidence of systematic breeding; most of the movements were to adjoining parishes, and most may not have involved sales. As in Brazil, most of the slaves who were moved interegionally were nonpredials, and the trade may have involved mainly a more efficient utilization of the skilled and domestic urban slaves.”°

Bride and Dowry Payments The other principal indirect means of acquiring slaves was through dowry and bride-payments. Slaves as bride-payments can be quickly dealt with, since this practice was confined to those preliterate tribes which circulated brides among groups against a countercirculating transfer of goods.°® Where slaves existed they sometimes constituted a part of the bride-payment. What is surprising, however, is that the practice was not as common as one might have suspected among slaveholding peoples with bridewealth marriages. The Dahomeans, for example, had elaborate bride-payments but there is no mention of slaves in the list of goods presented as payment.’’ Livestock tended to occur far more frequently as a standard item throughout Africa. Sometimes the high cost of slaves explained their absence in the bride-payment: among the Ibos the standard price of a female slave was one and a half times the bridewealth of the average maiden.*® There were cases, however, where the bride-price was considerably higher than the price of the slave, yet slaves were not mentioned as part of the bridewealth (for example, among the Duala of the Cameroons).”” For whatever reason, the use of slaves as bride-payments is attested in only a minority of slaveholding societies that had this custom. Among a few peoples it was an essential part of the bride-payment. For example, among

The Acquisition of Slaves 167

the Mende, slaves “formed an invariable and important part of bridewealth.” In view of the fact that the bridewealth was high and slavery very important, a considerable number of slaves must have been acquired in this way.” Among the Tuareg, slaves—especially newly acquired ones—were often part of the bridewealth of their masters;°' and among the predominantly patrilineal Wolof of Senegambia, it is curious that slaves acquired as bridewealth were inherited matrilineally, whereas bought slaves were inherited patrilineally.°

Perhaps the most dramatic case of the acquisition of slaves through bridewealth was that of the Nkundu of the Kongo, a primitive tribe among whom slavery was traditionally indispensable, not for economic reasons but because no marriage was truly legitimate until the bride-price was completed by the payment of the bosongo (which comprised, on average, two

slave women, but in some cases as many as five to ten slaves).°° “The strength of the marriage,” goes a traditional Nkundu saying, “is to be found in the slave.” It is also said that “by means of the bosongo the woman completely becomes the wife.” The bosongo of the Nkundu is one of the most curious reasons in the history of the institution for the emergence of slavery on a significant scale. Far more widespread was the practice of including slaves in the bridal dowry. The dowry, because it is found mainly among advanced peoples, figures as a mode of acquisition among all civilized slaveholding societies. In the ancient Near Eastern societies it must have been a significant means of acquiring slaves; it was the custom of the wealthy, especially in Babylon, to include several slaves in the dowry of their daughters. Slaves were included in the dowry in pharaonic Egypt; in India from ancient to modern times; among all the slaveholding peoples of ancient, medieval, and early modern Europe; and among all the slaveholding societies of the Americas.” As with bride-payments, some very odd customs occasionally surround

the use of slaves as dowry. In ancient Rome, being given as a dowry or

find this curious law:

pledge automatically made a man a slave.°’ Among the early Icelanders we A free woman engaged to be married was considered free from physical and other defects if she would bring a price not lower than that of a female thrall. But if it was found that she had such defects as would lower the price were she a female thrall, the punishment for the one who, with “knowledge of her defects, betrothed her to a man,” was outlawry in the first degree.”

Slaves as Money The acquisition of slaves as bride-payment or dowry is closely related to the use of slaves as money, and this constitutes yet another way in which slaves were acquired. Money, as is well known, has several functions: it is a unit of

168 Slavery as an Institutional Process accounting or a standard of value, a method of payment, a medium of exchange, and a means of storing wealth. In primitive and archaic economies, as Karl Polanyi and his students have emphasized, “the various functions of money are institutionalized separately” —that is, one kind of object could be

used as a unit of value, another for making payments, and so on. Multi-

functional money is a very modern phenomenon.” | | The interesting thing about slaves is that in many primitive and archaic societies they constituted the closest approximation to modern multifunctional money. In the ancient Near East, slaves were sometimes used instead of metal as a standard of value and a medium of payment for (among other things) brides, houses, and fines.°* In Burmese society until a century ago, slaves “were the currency in which a husband was compensated for the violation of his wife—two slaves for a poor but free woman, four for the wife of a merchant, eight for that of a rich man, fifteen for that of a lesser mandarin, and so forth.””” In both pagan and Muslim Africa, slaves were often used as money. Among the Mende of West Africa, slaves were exchanged for bags of salt and cattle: “A single slave was worth from three to six cows and a man, woman or child, were all considered as one ‘head’ of money. This was equivalent, later in the century, in 1890, to 3 pounds sterling.”’° Slaves were simi-

larly used in Yorubaland and parts of Central Africa.’' Among Muslim traders slaves were commonly “a store of value, albeit one which medical hazards made extremely risky,” as well as units of value and a common form of payment, especially of debts and fines.” It was in early medieval Ireland and Iceland, however, that we find the most complex employment of slaves as money. In Ireland the cumal, or female slave, was the highest unit of value.”” A cumal was equivalent to 6 to 8 seoit, and a set was worth between 3 and 8 cows. A cumal was also equivalent to 3 ounces of silver. It was the standard unit of value for fines. Thus in homicide cases the value of the life of a free man (the eric fine) was reckoned at 7 cumala. It perhaps is significant, however, that the honor price was reckoned and paid in other forms of money. The cumal, in addition, was used as a measure of the value of land. There has been some uncertainty in the past concerning whether the cumal was used as a method of payment, but Marilyn Gerriets has definitively established that the cumal did also serve this function of money.’* Interestingly, the mug, or male slave, was never used as a unit of value, but was sometimes utilized as a form of payment. The use of the female slave as a form of money during medieval times strongly suggests that she played an important economic role in early preChristian Ireland, but declined in economic significance by the pre-Norman period, her role as money being a vestige of her former economic value. With values of half a cumal frequently attested in the texts, it may be wondered just how half a human being was possible. There was no problem,

The Acquisition of Slaves 169 of course, where the cumal was being used only as a unit of value: half a cumal would simply be interpreted in terms of, say, 1% ounces of silver, or between a cow and a calf and 4 cows. But how was the problem handled if the cumal was also serving as a form of payment? A simple and elegant ex-

planation has been offered.’” In keeping with the pattern found in other tribal systems where slavery existed, the enslavement of women was rarely lifelong. The Irish therefore also reckoned the value of a cumal, in substantive terms, at 7 years labor by an able-bodied woman. Hence half a cumal would mean the payment of 3% years labor by an able-bodied woman. We have less information on early Iceland, but it is well established that female slaves were used as both units of value and forms of payment, and that, as in Ireland, silver and cows were also used as units of value. The equivalences were very precisely worked out by the Icelanders and did not differ significantly from those of the Irish. Thus: “I very strong and big male thrall = 24 aurar = 24 cows; | average male thrall = 12 aurar = 12 cows; | female thrall (for concubine) = 12 aurar = 12 cows; | female thrall = 8 aurar = 8 cows.””®

It may seem extraordinary that a human being should be used as money, and inhuman that the value she defined was only 8 cows. But in the comparative annals of slavery the Irish and the Icelanders placed a very high value indeed on their female slaves—as did the Danes, among whom an enslaved nun during the ninth century was valued at a horse with its gear, “not cheap,” as Eric Oxenstierna correctly comments.’’ In sixteenth-century Burma, 40 Indian slaves was the going price for a horse.’* In 1870, a normal year at the Kuka market on Lake Chad, a young adult man was worth the same as a good riding horse, although a young girl was worth a little more and a eunuch twice as much. However, in the glutted Bagirmi market of the Sudan two years later, women were being sold for 5 dollars; not long after, a cow was valued at 10 slaves, a young male slave at 6 chickens.”” And as late

as the third decade of the present century in the unadministered parts of northern Burma, especially in the triangular region, “slaves were so plentiful as to be worth no more than a few pigs.”*° IN THE LAST three chapters I have employed several analytic categories and reported a number of new findings. The most fundamental distinction is between the means of enslavement and the means of acquisition. Past failures either to recognize or to consistently apply this distinction have led to considerable confusion regarding the sources and distribution of slaves. Many eminent and otherwise very cautious scholars, for example, have discussed

the subject under categories such as “warfare,” “trade,” “kidnapping,” “birth,” and the like, without any apparent awareness of the nonexclusiveness of these categories. Invariably, discussions either imply (erroneously) that persons enslaved by means of warfare were directly acquired by the

170 Slavery as an Institutional Process slaveholders of the victorious society, or, more frequently, they ignore the issue of the disposal of enslaved captives as noncontroversial. With the aid of another distinction, between original and current means of enslavement, I have explored the relative contribution of prisoners of war and found that while enslavement by this means was generally important as an original means, it was unusual for it to rank as the most important current means of enslavement. For all but a small subset of significant slaveholding societies, birth was by far the major current means of enslavement. In this regard, the distinction between the social and biological reproductivity of a slave population has proved to be a useful analytic aid. The most important substantive contribution of these three chapters has been the identification and illustration of the seven major rules of inheritance of slave status. No doubt future research will refine and add to these rules. Since no one previously has attempted to identify and systematically discuss them, it would be surprising if, in this first effort, all the important types have been exhausted and all the relevant issues explored. In analyzing the means of acquisition, I have distinguished between direct and indirect means. Examination of the direct means has involved a reanalysis of the data on warfare and captivity. Contrary to conventional scholarly and popular opinion, my finding is that enslavement was not the normal fate of the vast majority of captives, even when captured or conquered by armies from societies with highly developed slave systems and a persistent demand for slaves. Slaughter, ransom, temporary imprisonment, colonization, impressment, and simple release were all at various times, separately or together, the more common fate of captives. Even during those special periods of history where most slaves in a growing large-scale slave society were enslaved as a result of captivity, most captives still did not suffer this fate. The vast majority of persons conquered or captured in battle by Greek armies during the classical period, or by Roman armies during the militaristic triumphs of the last three centuries before Christ, did not suffer enslavement. And the same was true of those who fell before the conquering armies of Islam. Of special importance is my finding that it was unusual for a conquering group to attempt to enslave on a mass scale, on their home territory, the free members of a conquered population. Where such attempts were made, the long-term consequences were always disastrous. Trade, we have shown, was next to birth the most important means of acquiring slaves. The other indirect means (bride and dowry payments and the use of slaves as money), though widespread, were minor compared with trading. A review of the five major slave-trading systems in human history disclosed many surprising parallels and historical continuities. What stands out most strikingly was the extraordinary centrality of the Mediterranean. This sea, with the nations that surround it, has consistently played a critical

The Acquisition of Slaves 171 role in slave-trading systems, sometimes throughout their entire course.”! The Mediterranean, central to the development of human civilization and lovingly celebrated in Euro-American historiography,” from the viewpoint of human oppression has been a veritable vortex of horror for all mankind, especially for the Slavic and African peoples. The relationship was in no

way accidental. ,

The

e.° Condition

of Slavery

HAVING EXAMINED how human beings were enslaved and

acquired, we need now to analyze their fate. We want to know how slaves adjusted to their masters and to their new condition, and how masters used their power in the relationship with their slaves. We must inquire too into the.way this relationship was accepted by the society at large. In short, we need to look at. what factors determined the adjustment, treatment, and institutionalization of slaves. The adjustment of the slave to his condition involved two basic kinds of relationships: that between master and slave and that between slave and community. Mediating these two relationships was a third, that between master and community. How the master treated the slave and how the slave responded depended, first, on factors intrinsic to the master-slave relationship.

We may see these as the private determinants of the relationship. But the master, however independent he may have wished to be in his relations with his slave, needed his community to both confirm and support his power. The community, through its agents, wanted this support reciprocated, if only to safeguard the interests of its members. These constituted the public determinants of the master-slave relationship. Some sprang wholly from the nature of the community, independent of the master’s position. Yet the relationship between the master and his community was never a static one. The master 172

The Condition of Slavery 173 wanted to influence public attitudes and deflect attempts to interfere with his

proprietary claims on his slaves. His ability to do so depended upon his power and influence; this itself was partly determined by the newfound importance he derived from possessing slaves. Nor was the slave a wholly passive entity. He might, in relative terms, be powerless; but he always had some choice. He might react psychologically, play the slave, act dumb, exasperate. He might lie or steal. He might run away. He might injure or kill others, including his own master. Or he might

engage in armed revolt. Barring all these, he might destroy his master’s property by destroying himself. To be sure, only a small minority of slaves ever made such drastic choices. Most chose simply to behave with selfrespect and do the best they could under the circumstances. Nevertheless, I know of no slaveholding societies in which some slaves at.some times did not rebel in some manner. Even where the slave remained completely docile, the very totality of his master’s power over him made his master dependent on him. While he might not have been a person of value, he was a thing of value—perhaps the only thing over which the master had true power. The master’s whole manhood might have been invested in him. Parasitic domination was a real possibility wherever slavery existed, as we shall see in a later chapter. LET US CONSIDER FIRST the factors that were intrinsic to the master-slave re-

lationship. The most important of these was the use to which the slave was put. Slaves, of course, have performed every known task. However, there was usually a primary use for which they were acquired, and the condition of the slave was to a large extent determined by this purpose: they might

have been acquired for prestige, political, administrative, ritual, sexual, marital, or economic reasons. It should be obvious that if slaves were acquired as secondary wives, concubines, or homosexual lovers, their material comfort (if not their peace of mind) generally would have been better than those acquired to perform agricultural or mining jobs. The primary use of the slave also determined whether he or she would be permitted to marry and rear a family. Whereas the denial of custodial power was an invariant

attribute of the slave condition, in practical terms there was considerable leeway in the degree to which slave families were allowed to remain stable. There was little variation among slaveholding societies with respect to the sexual claims and powers of masters over female slaves: I know of no slaveholding society in which a master, when so inclined, could not exact sexual services from his female slaves. What did vary considerably was the protection of slave women, and of slave unions, from interference by third parties. Where used as workers, the way in which slaves were integrated into the work force was another critical factor in determining their treatment. The

fate of slaves on highly regulated latifundia or plantation-type farms dif-

174 ~~ Slavery as an Institutional Process

fered radically from the lot of those incorporated as tenant farmers, and both differed from the kind of treatment experienced by household slaves or slaves of small family farmers. Another important determinant was the mode of acquisition. In almost all slaveholding societies a distinction was made between the slave bought as

an adult and the slave born in the household (or acquired as a child and brought up in the master’s household). Ties of sentiment usually developed between the master (and his family) and the housebred slave. The uses to which slaves were put interacted in important ways with their mode of acquisition in determining their condition. This was particularly the case in advanced precapitalist slave systems. Thus in ninth-century Iraq there were many slaves by birth who served as menials in the households of their lords and were treated with indifference, while the slaves acquired as adolescents or even as young adults and trained as soldiers became the trusted aides and confidants of their masters. This was true of all the Islamic slaveholding societies, as we have already seen, but it was equally true of other areas. The king’s slaves in Thailand; the eunuchs and other officials of many Roman, Byzantine, Chinese, and African courts; and the Greek tutors and clerks of Roman households are only a few examples of slaves acquired as adults who were much better treated and more highly regarded than slaves born in the household. To take a celebrated example from the private life of Cicero, we learn from Susan Treggiari that slaves or freedmen such as his secretary Spintharus, his accountant Hilarus, and his letter carriers Aegypta and Phaetho were all warmly regarded by their master and, later, patron; while he demanded “loyalty, affection, and not uncommonly self-sacrifice” from them, he in turn “treated his /iberti boni with some consideration for their own claims, with gratitude for their hard work and devotion, and with genuine affection.”' Educated slaves who were useful to Cicero were all manumitted. At the same time Treggiari shows that “Cicero and other authorities rarely name the humbler slaves” and may not even have known their names, although several of them were born in his household. The litter bearers who remained loyal up to the moment of Cicero’s assassination, the groom, cook, and maid of all work remain anonymous, and the only case of a laborer who was emancipated “is contemptuously dismissed as ‘operarius homo’ after he

played truant from his work, and 1s not named.’ , Closely related to these two variables is a third: the residence of the slave. It cannot be assumed that all slaves by birth were based in the master’s household, any more than it can be assumed that all bought slaves were

lodged outside the household. Both assumptions tend to be true of small, kin-based societies, but even among tribal pastoralists and many chiefdoms, and certainly in more advanced societies, many slaves acquired at birth grew up in separate quarters. In pastoral societies slaves were often segregated in

, The Condition of Slavery — 175 Slave villages where they specialized in agriculture, while other slaves, many of them bought as adults, were kept as retainers in the household. This pattern held among all the early states of the Sudan, as well as the emirates of

northern Nigeria, right up to the end of the nineteenth century.° While it is certainly true that slaves of the household, whether born in the household or not, tended to assimilate their master’s culture at a faster rate than those quartered elsewhere, it must not be assumed that the household slave was necessarily better treated than the tenant or field slave. Nonetheless, being in the household had the advantage of proximity to the comforts of the master, and the slave who won the master’s favor was indeed privileged. In favor of the house-born slave too was the fact that in most precapitalist slave societies masters were reluctant to sell such slaves, hence they were spared one of the major risks of slavery, that of sudden and arbitrary disruption in their lives. Among the Mende, for example, it was considered shameful to sell a house slave.* Sanctions against such sales were not

formalized among the more commercially advanced peoples. However, while an unscrupulous or financially embarrassed master could sell houseborn slaves within his society, among a number of such advanced precapitalist cultures there were legal prohibitions on the sale of slaves abroad, especially house-born slaves. In both ancient Mesopotamia and Palestine

slaves could not be sold to foreigners or gentiles. And in Egypt during Roman times, where we find considerable data on the oikogeneis, or houseborn slave, the sale for export of house-born children of slaves incurred severe legal penalties.” Proximity to the master also carried enormous risks and disadvantages. The slave was under the constant supervision of the master and therefore subjected to greater and more capricious punishment and humiliation than those housed elsewhere. This was particularly true of the female slave, who in every slaveholding society, from the most primitive to the most advanced, ran the additional risk of the jealousy and vengeance of the “free” women of the household, especially the master’s senior wife. The famous adage should have run: hell has no fury like a free wife scorned in favor of a slave. In contrast, the slave who lived apart, while materially more insecure and more exposed to the vindictiveness of free third parties, had a much greater measure of independence. Most slaves in most societies valued this partial “freedom” far more than the dubious material delights of the great house. There is abundant evidence from ancient Greece, Rome, and elsewhere that the condition most coveted by the slave was to be able to live on his own and hire himself out or otherwise provide for himself. A fourth factor bearing on the condition of the slave was his original means of enslavement. Its direct influence, however, was weak. It was of far greater importance in its impact on the public determinants of the slave’s condition, that is, in its effect on how the society at large responded to the

176 Slavery as an Institutional Process slave and permitted his adjustment to public life and communal activities. A captive from a traditional enemy would clearly have more problems with the community at large than, say, a local person who fell into slavery as a result of destitution. Still, this cut both ways. In several societies, for example medieval Wales and the Toradjas of the central Celebes, the captive was regarded with less disrespect than the local person who had fallen into slavery, because the former was less responsible for his condition. _ Another cluster of variables influencing the condition of the slave relates to his personal characteristics. To begin with race, we note that there were fifty-five societies in the world sample on which adequate data were available. Of these 75 percent had populations in which both slaves and masters were of the same mutually perceived racial group, 21 percent had populations in which masters and slaves were of different racial groups, and 4 percent had populations in which some slaves were of the same racial group as their masters while others were not. It has often been remarked that slavery in the Americas is unique in the primary role of race as a factor in determining the condition and treatment of slaves. This statement betrays an appalling ignorance of the comparative data on slave societies. A great deal, of course, depends on what one means by race. I take the racial factor to mean the assumption of innate differences based on real or imagined physical or other characteristics. In these terms, there have been numerous slaveholding societies where race was socially important; it is not at all obvious, though, how race influenced the condition of slaves. Throughout the Islamic world, for instance, race was a vital issue. The light-skinned Tuareg and related groups had decidedly racist attitudes toward the Negroes they conquered.° Throughout the Islamic empires, European and Turkish slaves were treated quite differently from slaves south of the Sahara Desert.’ In pre-European Malaysia, the Arab rulers viewed the conquered native population with utter contempt.* In Han and later China, the darker and physically distinct border peoples were considered racially

inferior to the heartland Chinese; not only were they considered natural slaves, but their harsh treatment was tolerated whereas such treatment of a truly Chinese slave would be punishable by law.’ The border peoples, however, responded in kind. As late as the early twentieth century the Lolos of Taliang Shan kidnapped and enslaved Han Chinese even though they were under the sovereignty of China. In a remarkable inversion of racial stereotype the Lolos called their upper-class members “black Lolos” and the subject population “white Lolos.” White skin, so highly prized among the Han Chinese, was despised by the swarthy black Lolos and became a mark of ser-

vility and a way of identifying the Han slaves.'° a 7

Among the medieval Scandinavians, where blond hair, white skin, and blue eyes were the somatic ideal, the stereotype of the thrall was so consistently different that one is tempted to think that most of their slaves came

The Condition of Slavery 177 from the darker European or even Asian peoples (although we know that this was not the case): “The thralls were said to be ugly. The thrall mentioned in the Rigspula had swarthy skin, a hideous countenance, and a flat nose.”'' Foreign slaves who conformed to the Icelandic racial ideal were more favorably regarded. This, for example, was true of Freystein, a thrall belonging to Thorkel Gaetisson: “He was neither ugly nor hard to manage like other thralls, but, rather, gentle and of good manners and more handsome than almost anyone else, wherefore he was called Freystein the Fair.” More important than the color of their skin, however, was the belief that thralls as a group were innately inferior in mental qualities and other personal attributes. This is well illustrated by the story of Hjor, the Norwegian king, and his wife Ljufvina, a Russian princess whom he had captured.

Ljufvina gave birth to twins but did not like them because they were swarthy. She therefore exchanged them for Lief, a lighter-skinned child of her bondswoman by a thrall. The king, however, disliked Lief, finding him lacking in manliness. One day the queen asked a poet to evaluate the three children as they played together. He immediately recognized the nobility of the twins and the slave origins of Lief, whereupon the queen took back her twins. '”

It is not true either, as is so often claimed, that race was not an issue in the classical world. Although it was certainly of less importance than in the modern Americas or parts of the Islamic world, it did operate in a fairly significant way. We have already observed that in Greece and Rome slaves came from almost every racial group; and “Ethiopians,” as most Africans were called, were very much present. Nor should one underestimate the somatic differences between the lighter, blue-eyed, northern slaves and their darker, curly-haired, brown-eyed, Roman masters. The important question, of course, is not the mere presence of physical differences but their sociological significance for Greeks and Romans. Certainly, when one considers the striking emphasis the Greeks placed on physical beauty in both sexes, especially males, it would be sociologically unlikely that somatic factors did not figure in their treatment of slaves. A beautiful young boy slave who came

close to the Greek physical ideal would almost certainly end up as the homosexual lover of his master. Physically attractive women and less “virginal’ but still pretty boys ran a far greater risk than their less handsome counterparts of being forced into the lucrative prostitution trade in which many Greek and Roman masters engaged.'” Those who, to Greco-Roman eyes, were less physically attractive races, were spared such fates; but aversion to them may well have been expressed in other ways. Frank M. Snowden, Jr., is right in claiming that the racial factor weighed less heavily for blacks in antiquity than it did for those in the Americas, but he exaggerates when he suggests that there was little racial prejudice.'* There is not much literary evidence to go on, but if the occa-

178 Slavery as an Institutional Process sional asides of the satirist Juvenal are reliable, strong racial antipathy was not uncommon in Rome." From the portrayals of blacks in sculpture and painting with their features persistently caricatured as bestial and grotesque, it seems safe to conclude that Negro features were not an asset in the slaveholding societies of the Greco-Roman world.'® Racial difference from the master did not, of itself, always work against the slave. Among the Arabs and Ottoman Turks white slaves were specially prized.'’ Men as well as women of different races were recruited for sexual purposes.'* In nineteenth-century Egypt, Caucasian slave women were the most prized members of the harems of the upper classes.'’ Nor was a black

skin always a disadvantage. In imperial China and Islamic India, black slaves were valued for their exotic appearance and fetched the highest prices.*” And young black pageboys were the rage in the boudoirs of eighteenth-century England and France, although what they did with their leisured mistresses as they grew older is best left to the imagination.”’

More important than race, and more widespread in its repercussions, was the ethnic difference between master and slave. Through its interaction with this variable the means-of-enslavement factor could be most effective: prisoners of war were more apt to be of a different ethnic and religious group

than slaves originally obtained by other means. But this was only a tendency: wars were often fought between peoples who are ethnically close. This raises the interesting question of the reluctance of peoples to enslave those with whom they shared a common culture and felt a sense of ethnic identity. From the comparative data on this problem we may draw two conclusions. First, it was simply not the case that slavery within the ethnic group never existed—that, as Henri Lévi-Bruhl claimed, endoservitude was an impossibility.” Slaves were recruited from within the community in a Significant number of slaveholding societies. There was, nonetheless, a universal reluctance to enslave members of one’s own community, hence the need to redefine them as outsiders. Yet the ethnic group is often wider than the community or the state, and where such cross-societal ethnicity existed it is useful to inquire whether there was a reluctance to enslave fellow ethnics. The data suggest that while there was indeed such a reluctance, different groups responded to the dilemma in different ways. The most common response was either to kill or to ransom or sell elsewhere-enslaved fellow ethnics, but not to enslave them. Upper-class black Lolos were always killed when captured by fellow Lolos in their intraethnic skirmishes.” The agricultural Vai of West Africa were constantly at war with one another. Before

the European demand for slaves on the coast, Vai captives were always either killed or ransomed, and only the defeated group’s slaves or non-Vai subjects were taken into slavery. With the European demand for slaves on the coast, Vai captives routinely were sold there.** Among the Tuareg there

was a carefully observed agreement that fellow Tuaregs when captured

, The Condition of Slavery 179 would either be released or ransomed and only Negro captives would be enslaved.”’ The ancient Greeks clearly agonized over the issue and were generally loath to enslave fellow Greeks from other states, but where Greeks

from hostile states were taken in war and were not ransomed, they were enslaved.

What all these choices amounted to was a distinct class bias in the enslavement of fellow ethnics. Upper-class and wealthy persons from hostile states, tribes, or clans were usually ransomed or killed. The fate of lowerclass or less prosperous persons was more varied: where an external market existed, they were sold away from the region; otherwise they were either killed (especially among less advanced peoples) or were reluctantly enslaved. Among classless, acephalous, slavehoiding peoples intraethnic slavery was rarely prohibited.”° While the ethnic factor was important in determining who became enslaved, it had surprisingly little influence on the treatment of slaves. Data were obtained on fifty-seven of the slaveholding societies in the sample of world cultures. In 75.4 percent slaves and masters were of different ethnic or tribal groups; in 15.8 percent the two were of the same ethnic group; and in 8.8 percent some slaves were from the same group as their masters, while others were from other ethnic groups. Ethnicity did not significantly correlate with any of the variables that attempt to measure the treatment and condition of slaves. Another attribute of slaves that influenced their condition was gender. It should not be assumed that female slaves were always acquired primarily for sexual purposes. Among most of the more developed slaveholding societies of Africa, women—both free and slave—played a major role in food production. Even where the traditional female role was minor, slave women were utilized as farmers. For this reason sex was not as critical a factor as might be imagined, and the sex ratio of the slave population related to the overall treatment of slaves in a wholly unexpected way, as we shall see. Regardless of treatment, women were more easily assimilated into the community than men, for reasons examined earlier. —

Skill also played a part in determining the condition of the slave. The captive who possessed skills that were in short supply in his master’s community, or the slave of the house who was trained into such skills, was obviously more valued by his master and was likely to have been much better treated. In many slaveholding societies slaves were acquired with the specific purpose of introducing skills to the slaveholding group. Slaves who knew how to carve totem poles, for example, were greatly prized among the

Tlingit and were often hired out to individuals who were indifferent carvers.*’ Greek society came to rely heavily on skilled craftsmen for its urban industries, and this fact, more than any other, determined the character of Greek slavery (not to mention its overall economy).”* The same was

180 Slavery as an Institutional Process even more true of Rome, where skilled and literate slaves came to dominate not only urban industries, but education, the arts, theater, and literature.”’

And perhaps to an even greater degree the same held for the Islamic empires, especially in their early periods. According to Samuel S. Haas, slaves were “the leading elements” in the cultural transformation of the Islamic peoples between the period of Muhammad and the fourteenth century. Apart from their role in politics and warfare, slaves “exerted strong in-

fluence in the realms of public administration, religion, arts and crafts, music, poetry, grammar and learning in general.””*°

Many of the striking similarities between Greco-Roman and Islamic slavery derive from the reliance on skilled slaves—features such as the high incidence of manumission, the urban character of the institution, and the absence of any strong tendency to impose caste attributes on the slave or freedman populations. Even where slaves were not a dominant part of the economy, the skills they introduced could have major implications for the particular culture.

Toward the end of the fifteenth century, as an example, many new technologies were introduced to Europe via Italy by Asian slaves, including the vertical-axle windmill, the hot-air turbine, and a new type of governor.”!

While the possession of a skill in great demand usually worked to the benefit of the slave, this was not always the case. Specialists of various sorts in Africa and other parts of the precapitalist world were often despised and feared and constituted an outcaste group. Slaves with such skills, while valued economically, would naturally suffer the special contempt shared by all those who practiced their craft.*” Another important variable determining the treatment of slaves was the relative size of the slave population. The relationship is both interesting and complex and will be discussed at some length shortly. Two other independent variables are worth mentioning, to conclude this portion of the discussion. One is the level of absenteeism among slave owners. My earlier study of slavery in Jamaica showed that one of the major factors determining the condition of the slaves in this large-scale system was the high level of absenteeism among masters with large holdings.*’ In exploring whether the same held for precapitalist slave systems, I found, first, that genuine absenteeism existed in only a small proportion of all slave societies: 6 percent of those in the Murdock sample of world cultures, and no

more than 10 percent of all societies with significant levels of slavery. Among the economically less advanced peoples, genuine absenteeism existed mainly among pastoralists such as the Somali, certain Arab tribes, and the Manchu.“ It is important to distinguish between absenteeism proper and the simple living apart of the master class, which we find in those situations where slaves were employed as tenant farmers. A considerable number of small-

The Condition of Slavery 181 scale precapitalist societies utilized slaves in such tenant settlements: most of the Germanic tribes who kept slaves,” and many African groups such as the Tuareg,*° the Ashanti,*’ the Sherbro,*® and the Mende. What Kenneth Little

says of these satellite slave villages among the Mende holds for all the others, that “legally speaking the slaves could be described as tenant settlers.”°? The Hausa were a borderline case, with their vast slave villages some three thousand strong. Most of these comprised enslaved tenant farmers, but several were organized along more formal lines, with direct supervision that brought them closer to the latifundic pattern.“° I. M. Diakanoff’s

term, “pseudo-latifundic,” used to describe some kinds of large farms manned by enslaved tenant farmers in the ancient Near East, is highly suggestive though problematic.”!

Whether the owner was a genuine absentee or not, the tenant-farming arrangement usually worked to the benefit of the slaves; they controlled the means of production, usually organized their own schedule of work, and had a fair degree of autonomy in conducting their personal lives. The impact of absenteeism only became significant when it occurred in combination with a latifundic or plantation type of socioeconomic organization, or with largescale mining. Although the combination was infrequent, when it arose, the consequences were invariably disastrous for the lives of the slaves. The most notable cases in the precapitalist world were the Laurium mines of Athens during the fifth century B.c.; southern Italy and Sicily during the period of

the late republic and early empire; the slave system of the dead lands of lower Iraq during the Abbasid caliphate, especially during the late eighth and early ninth century a.D.; many of the slave estates of Visigothic and Muslim Spain; the Italian slave plantations on the Mediterranean islands during the late medieval and early modern periods; the royal slave plantations of nineteenth-century Dahomey, West Africa; the Arab-owned absentee slave farms of nineteenth-century coastal East Africa; and in the Orient the large slave estates of the city-based slave owners of Korea during the Koryo and, to a lesser extent, the early Yi periods. It is in these and the slave plantations of the Indian Ocean colonies, the Banda Islands of Indonesia,

the Mascarene Islands of the Indian Ocean, eighteenth-century South Africa, and the modern Americas that the spiritual, social, and material condition of the slaves reached its lowest level. Finally, a negative finding is worthy of note, if only because it was given such prominence in the work of the early comparative students of slavery, especially H. J. Nieboer. From the emphasis placed on the classification of societies according to their dominant mode of subsistence (hunter-gatherers, fishers, pastoralists, agriculturalists) one would believe this variable to be important in accounting for variations in the treatment of slaves. My own analysis suggests that mode of subsistence had no effect whatever on the way in which slaves were treated.

182 Slavery as an Institutional Process LET US NOW EXAMINE more precisely the influence of these independent variables on a select number of factors that reflect the treatment of slaves by their masters and other nonslave persons.

The Peculium In all slaveholding societies the master had nearly exclusive proprietary claims and powers over the person, labor, possessions, and progeny of his slaves. The fundamental feature of slavery, in law, was the fact that the slave

could not be a proprietor: he or she was, quintessentially, a property-less person. From this fundamental disability flowed, both in legal and in socioeconomic terms, all che other manifold disabilities of the slave. However, it is also true that in all slaveholding societies the slave was allowed a peculium. The peculium may be defined as the investment by the master of a partial, and temporary, capacity in his slave to possess and enjoy a given range of goods. The peculium differed from genuine property in that, first, it never included ai/ the proprietary capacities. The master always reserved a claim on the possessions of the slave. The slave, however, was allowed the usufruct of the possessions in question and could exercise certain powers and privileges over them in his relation to all third parties. In rare cases such powers and privileges may even have extended to the master, although never to the point of denying the master’s ultimate claim on the possessions. The fact that the master reserved this ultimate claim meant, in theory, that the slave was not permitted to dispose of the peculium. Practice invariably followed legal precept here. Yet masters often gave their slaves permission to dispose of the less important movables in the peculium, as well as some of the income generated from any capital assets. Slaves were also frequently allowed to trade and engage in business, using their peculium if the objective was to enlarge it. The peculium was always a temporary possession. This too was implied in the claim on the property reserved by masters in all slaveholding societies. The usufruct could be withdrawn at any time, but the nearly universal tendency was for the slave to enjoy it for the course of his lifetime. On his death

the possessions reverted to the master, who reassigned the peculium to whichever slave he wished. Usually it was in his best interest to reassign it to

a descendant or kinsman of the slave, but I know of no slave society in which slaves had a recognized power or privilege to endow legacies or to inherit them.

There was considerable variation among slaveholding societies in the objects over which possession might be given a lifetime usufruct. Where slaves worked the land in allotted parcels, it was understood that this was only for their keep and for the benefit of their masters. They were usually allowed to include in the peculium what remained from the produce of the

The Condition of Slavery = 183

land after they had provided for themselves and their masters, but the land itself was almost never considered theirs. Of course, in most lineage-based societies land itself was not part of the exclusive property of a single individual: even nonslave persons had only a usufructory claim on it. The difference between the slave and the nonslave was that the nonslave was usually entitled by birth to such a usufruct, whereas the slave used it only at the pleasure of, and in the interest of, his master.** Occasionally one finds what appear to be exceptions to the general prohibition of the inclusion of land in the peculium. The two most striking cases

are Korea during the Koryo period and Russia during the sixteenth and early seventeenth centuries. Closer inspections of these two cases, however, raise questions about whether they were indeed exceptions. Ellen Unruh, who discusses the Korean case,” tells us that “by mid-Koryo, slaves appear to have been permitted to own and dispose of property, i.e. land,” and she rightly finds the practice “unusual.” She argues that to assume that the own-

ership of land “negates the slaveness of the Korean slave” is to adopt an ethnocentrically Western view of both slavery and liberty. “Has materialism so warped our thinking,” she asks, “that liberty is interpreted as the liberty to make money and own property?” The important point, she observes, is

that “slaveness has a moral, not a material, connotation.” What was objected to in Korean law and practice was not the simple ownership of land but the ends to which the slave directed such ownership. If he used his possession to acquire the trappings of an honorable person and to deny his inherent baseness, then the law came down on him with all its force.“ Our suspicions are reinforced when we examine the second supposed exception. Richard Hellie observes that some slaves in Muscovy “were allowed to own other slaves, landed estates and urban property.”*” However, only a small minority of slaves—what he calls “elite slaves” —were so privileged (2.4 percent). These slaves constituted a managerial class for the landowning elite, and their existence was the direct result of the law code of 1550

that stewards had to be slaves. Hellie speculates that “they provided second-level managerial and other highly skilled talent in a society in which people of high status were unwilling to subordinate themselves to anyone else.” It is questionable whether these persons were, strictly speaking, slaves. Most of them were so-called registered slaves (dokladnoe). Hellie lists seven categories of slaves, only the first two of which were “hereditary” and “full”

slaves. By implication the registered slaves were neither, which automatically excludes them from the group of persons who are the object of this study. Most so-called slaves in Muscovy were “limited-service contract slaves” or indentured servants. Here is another transitional situation, in which the demands for a more flexible labor force at both the managerial and working-class levels were met by an adaptation of traditional patterns of bonded labor. The registered slaves of Muscovy were really bonded re-

184 Slavery as an Institutional Process tainers, a temporarily nonfree but strictly nonslave group substituting for the free managerial class which Russia, in its slow and painful transition to capitalism, was reluctant to encourage. Significantly, there is no mention of “full” or hereditary slaves owning land. We must conclude that, like Korea, a case has still to be made that this was an exception to the general tendency of slaveholding groups to deny slaves possessions in land as part of their peculium. The restriction on land was less of a problem for slaves in the urban sectors of the commercially more advanced slaveholding societies. Slaves in the

ancient Near East, in the Greco-Roman economies, all over the Islamic world, and in medieval Europe performed every kind of economic activity that free men engaged in—and often did much better financially than the latter. Gifted slaves sometimes earned vast sums as bankers and agents of their masters in all these societies, and in many of the city-states of nineteenth-century West Africa. In sixteenth-century Seville skilled blacks engaged in all the artisan crafts, and their competition was deeply feared by the city’s guilds.”°

While land was generally excluded from the peculium, slaves were almost always included. The servus vicarius (slave of a slave) was a universal occurrence. I know of no slave society in which slaves who could afford them were denied the purchase of other slaves. The fact may seem surprising at first, but on further reflection it ceases to be so. If slaves were the extension of a man’s person and honor, so were his slaves’ slaves. There could be few greater testimonies to a man’s power than the fact that even his slaves possessed slaves. At the same time, nothing more confirmed the loyal slaves’ acceptance of the condition of slavery and their own enslavement to the master than their willingness to own slaves themselves. The servus vicarius was the best way of making it clear to all that slavery was part of the natural order of things. Undoubtedly, too, the servus vicarius fulfilled for the slave a function more akin to the modern psychological usage of the term “vicarious”: the wealthy slave vicariously experienced the status of his master in his relationship with the slave. While this undoubtedly was good for the psychological relationship between the owner of the vicarius and his own master, it was not so good for the vicarius who, often as not, was very much the scapegoat. If we set aside such unusual cases as the vicarii of the imperial slaves of Rome and the Arab caliphates (to be considered in Chapter 11), no condition on earth was less enviable than that of the servus vicarius. In theory, it would seem that slaves should not be able to redeem themselves with earnings from their peculium: the master, after all, was being paid for one part of his property (his slave) with another (his slave’s peculium). In practice, most slaveholding societies found ways of getting around this nice legal problem.*’ Societies varied in their rules pertaining to the disposal of the peculium when the slave was sold or when he bought himself, although there was not

| The Condition of Slavery 185 much of a problem in practice. In many premodern societies the slave who accumulated a peculium worthy of the master’s notice was not likely to be a slave the master would want to sell. The surest sign anywhere of a “good and loyal slave” was a slave with a large peculium. And where the slave bought his own freedom, it usually cost him his entire peculium. If a slave had saved for years to purchase his freedom, it would hardly have been in his interest to save beyond the sum and time when he could redeem himself. In Rome, where unusual cases were most likely to occur, the slave’s peculium usually went with him when he was sold, “and so it would on manumission inter vivos unless expressly withheld.”** Recognition of the slave’s peculium was very nearly universal. Societies varied only in the degree to which the peculium was legally or socially sanctioned. In the ancient world the Cretan state of Gortyna is held by some to be unusual in legally sanctioning the peculium (although the interpretation of Gortyn laws and the extent to which they were actually applied is highly problematic).*” Less ambiguous was the West African case of Ashanti during the period reconstructed by Rattray (the late nineteenth and early twentieth centuries). He tells us: Slaves could and often did amass considerable wealth and attain to considerable power. A master encouraged his slave and helped him in every way to do so, because ultimately everything the slave possessed went to the master.

A master could not deprive his slave of his self-acquired property. An Ashanti proverb sums up the situation tersely, thus ... A slave may eat to repletion while his master remains hungry, but what the slave has, 1s, after all, only wind in his stomach.”

In an earlier work Rattray observes: “Lands were also granted to a favorite household slave for life with reversion to the donor, but the slave’s children, as long as they served the household, were often in practice allowed to continue to occupy and use the land after the death of the original grantee.””!

herit.”” ,

The Ashanti master in exceptional cases also allowed a slave child to inThe practice is easily explained in the case of the Ashanti. The society was matrilineal, but patrilineal blood ties (the ntoro principle) remained strong. A slave child by a slave woman had no formal matrilineal connections. Every “free” Ashanti dreaded the prospect of extinction of his line and his household. If the master had no matrilineal heirs except a slave child, he preferred to allow the slave to inherit over patrilineal blood relations; the slave, “having no abusa, or in other words no other home, when he came into an inheritance carried on the old master’s home at the same spot and the rites of honor of the departed spirit.” The universality of the peculium is not difficult to explain. It solved the most important problem of slave labor: the fact that it was given involuntarily. It was the best means of motivating the slave to perform efficiently on his

186 Slavery as an Institutional Process master’s behalf. It not only allowed the slave the vicarious enjoyment of the capacity he most lacked—that of owning property—but also held out the long-term hope of self-redemption for the most diligent slaves. The master lost nothing, since he maintained an ultimate claim on the peculium, and he had everything to gain. The ancients recognized this as much as the masters of all other large-scale slave systems. Varro, writing in the first century B.c.

in his treatise on agriculture, advised that “the [slave] foremen are to be made more zealous by rewards, and care must be taken that they have a bit of property [peculium] of their own, and mates from among their fellowslaves to bear them children; for by this means they are made more steady and more attached to the place. Thus, it is on account of such relationships that slave families of Epirus have the best reputation and bring the highest

price.””* | |

Manipulable statistical data on the peculium were obtained by separating the societies in the sample of world cultures into two groups: those in which the peculium was recognized, strongly sanctioned, and encouraged, and those where it was weakly sanctioned, had no status in law, and was not especially encouraged. Of the forty-four societies on which there were adequate data, 70.5 percent fell into the first group and 29.5 percent into the second. Cross-tabulation of this variable with others showed that the minority of societies that did not sanction or encourage the peculium were the most brutal in the overall treatment of slaves. Masters could kill slaves with impunity to a greater degree, and they were more inclined to provide for the slave rather than to allow the latter to provide for himself. Crimes against slaves either went unpunished or were punished to a far lesser degree than crimes against nonslaves. Neither race, ethnicity, nor size of the servile population relate in any significant way to attitude toward the peculium. There is, however, some association with the prevailing mode of subsistence. Fishing communities were most inclined to curtail the peculium, while pastoralists were most likely to recognize and encourage it. In fishing communities slaves customarily were used to perform menial tasks under the direct supervision of their masters. The basic resource, fish, was acquired without a great deal of effort or complex planning and execution. There was therefore little need to motivate the slave. In pastoral societies, on the other hand, slaves were usually kept as agricultural specialists and often left for long periods on their own. There was an urgent need to motivate them to work on their own, and the peculium with its prospect of eventual self-redemption was always the best way of doing so.

Marriage and Other Unions It has been claimed by an eminent student of comparative slavery that marriage is incompatible with chattel slavery.”’ I cannot fully agree with this

The Condition of Slavery 187 view. The compatibility of marriage and slavery depended on the nature. and sociological significance of wedlock in a given society as well as the number of marital arrangements permitted. If a society had only one kind of marriage, and the institution was defined in such a way that it automatically implied the legitimacy of progeny and the custodial powers of parents, then by definition marriage would be incompatible with slavery. Many societies in both the premodern and modern world, however, recognize a range of permissible unions between adults. Some of these may be quite consistent with slavery; others may not be. It is best to remember that, sociologically, marriage and the family are closely related but different institutions; the former regulates the sexual unions of adults, the latter provides the framework within which children are born and reared. Usually marriage legitimizes not just the cohabitation of the parents but the status of their children. Still, there are many notable exceptions among so-called free populations. In Islamic societies the children of concubines are “legal” children even though the parents remain unmarried, and the same is true in many other societies. In Jamaican and Puerto Rican law all children are legal even though most unions are illegal. Hence primary familial bonds may be initiated and are legal where there is no marriage among parents. And, conversely, it is possible for legally married parents to produce children who are not legally their own: this is exactly what happens in a matrilineal society where a person is the legal child of his mother’s brother and not of his biological father, in spite of the fact that the latter is legally married to his mother.” The range of options open to both masters and slaves was wide, and no simple assertion concerning the incompatibility of slavery and marriage is tenable. In 97 percent of the societies falling in the sample of world cultures, masters recognized the unions of slaves. In not a single case, however, did such recognition imply custodial powers over children. The Mende is one of the few lineage-based societies in which masters were found to discourage unions. According to John J. Grace: “Some owners were so afraid of the growth of family feeling among their nduwonga that they forbade a slave woman to have successive children by the same man.”°’ Even where marriages among slaves were not recognized, it was unusual to find such deliber-

ate discouragement of regular unions in the precapitalist world. The Laurium mines of Attica during the fifth century B.c., the slave latifundia of late republican Rome (although, as we learn from Varro, foremen were allowed to have common-law wives), and the dead lands of lower Mesopotamia in late ninth-century Iraq were notable—and exceptional—instances of this attitude among the more advanced of the precapitalist slave systems. In most lineage-based societies the slave either paid no bride-price, or the bride-price (usually well below that of a free woman) was paid by his master. Invariably the slave had to seek the permission of his master before taking a wife, especially when she belonged to another master. This was so even among the Ashanti, where slave marriage was strongly sanctioned.

188 Slavery as an Institutional Process The Islamic legal traditions are among the most liberal in the precapitalist world. Most authorities deny that an adult male slave could contract a marriage of his own volition, but if a slave was a Muslim he was considered

legally competent to marry after receiving his master’s permission. The master reserved the power to compulsorily marry off his male or female slaves in all Islamic traditions except that of the Malikis. In this tradition adult male slaves could marry of their own accord; however, the master reserved the power to ratify such unions and to terminate them by repudiation. Male slaves were allowed two wives in most traditions and were permitted to divorce them.”* Because slaves could not acquire the legal status of spouses, they could not in the full sense commit adultery or fornication. For this reason they were spared the death penalty in such offenses. At the same

time, fornication could not be committed against them. Spouses did not have custody over their children.” One should be careful not to idealize the Islamic situation. Even though it was better than most arrangements for slaves, it still left a great deal to be desired. In all Islamic societies the slave woman was at the sexual mercy of her master, in both law and practice. The law forbade masters to break up families, but where economic factors required it, masters found all sorts of ways of getting around such religious prohibitions. A few Islamic states even discarded the religious precepts altogether. The Somali master reserved and frequently exercised his power to sell the mother separately from her children, and of course “the morality of the slave women was not safeguarded

by any law.” Let us consider some typical cases from the non-Islamic world, moving as usual from less to more advanced societies. Among the Ashanti a male slave was allowed to pay the asida, or bride-price, and when he did so he could claim damages for adultery against anyone. What is more, if the master was the culprit, he paid twice the usual amount and was publicly humiliated for so debasing himself.°' Even so, the children of slaves belonged to the master unless the mother was a free woman. The laws of Hammurabi say a good deal about slaves, but almost nothing about the marriages of slaves to each other. The laws only touch the subject briefly in reference to the marriage of certain classes of official slaves to free women.®* Other evidence strongly suggests that family ties were not always respected and that although the sale of families as a unit was “not uncommon” there was nothing to prevent ruthless masters from selling spouses separately.* The documents on the Third Dynasty of Ur, as analyzed by Bernard J. Siegel, imply that while marriages between slaves were “commonplace,” the slave did not have the right to protest the sale of his or her child by the owner, “in other

words [that] he did not have the potestas over his own children.” In imperial China the slave family, in particular the slave wife, was protected against third parties but not against the economic or sexual demands

| The Condition of Slavery 189 of the unscrupulous master. Slaves were considered bound to each other by the laws of filial piety, and masters were encouraged not to separate them. It is revealing, however, that the slave family was typically much smaller than the average nonslave family and that slaves had no surnames. William Westermann asserts that slaves in Athens who lived apart could marry and found their own households, but this is sheer speculation;®’ moreover, such slaves constituted only a minority of all slaves. Roman legal theory, which probably did not depart much from practice, certainly influenced most subsequent slaveholding societies in the Western world. Slave marriages were not recognized in law; slaves could not establish connubium, only contubernium, and the master had the power to interfere as he saw fit. In Cato’s day there seems to have been relatively little respect for the contubernium, but by Varro’s day and thereafter it was more like a common-law marriage. As R. H. Barrow observes, by the early empire the “ius gentium [was] triumphing over the ius civile”’ and jurists were “as ready as the slaves themselves to speak of maritus, uxor, filius, pater within the boundaries of slavery.”°> The slave, however, never became a paterfamilias and could never exercise potestas. The triumph of Christianity did lead to a significant improvement in the marital and familial condition of slaves in the late empire and during the Middle Ages. A law of Constantine passed in A.D. 334 forbade the separation of slave families. Slave marriages were given religious sanction although not legal confirmation. In seventh-century English society if one of a Slave couple gained his or her freedom, the free party was allowed to buy the freedom of the one still enslaved or else marry a new partner.°’ According to Marc Bloch, religious validation of slave marriages was one of the important religious actions which “gave its aid to the general movement that transformed slavery.”’” One must be careful not to exaggerate, just the same. The church throughout the Middle Ages justified slavery as part of the law of man and the consequence of sin. While it required the baptism of slaves, it

sanctioned the sale of Christians (except to Jews and Muslims). It encouraged masters to respect the integrity of slave marriages and families, and one finds an occasional law that reinforced such entreaties. On the whole, though, the master remained supreme in his power over the marital and familial lives of his slaves. Throughout Europe an unscrupulous master could always sexually abuse his female slave, married or not; and at no time did the slave father have custody over his child.”’ Throughout the modern Americas the unions of slaves and the integrity of their households rarely received legal sanction. Nor was the church any more effective in this regard than it was during the late ancient and middle ages. (There were exceptional cases, such as portions of nineteenth-century Brazil.) The actual stability of unions and households varied with the kind of slave economy, with the demand and supply of slaves, and with the sex

190 = Slavery as an Institutional Process

ratios of both the slave and free populations, issues we have already discussed. Where the plantation economy was dominant, demand for slaves high, external supplies available, and males in the majority among both masters and slaves, slave unions and households tended to be highly unstable (the U.S. South before 1808 being an important exception). This was true of the French and British Caribbean and many sections of Brazil up to the last quarter of the eighteenth century. In cases where the plantation system was dominant and the demand for slaves high, but external supplies were either curtailed or cut off, the resulting high cost of slaves made their natural reproduction both profitable and necessary. For these reasons stable unions and households were encouraged,

sometimes even required, by the master class. The U.S. South during the nineteenth century is the best-known case. However, the same situation occurred in the British and French Caribbean during the last decades of the eighteenth century and in the early nineteenth century”* Finally, where the plantation system was not dominant and external supplies were available and kept up with demand, the unions and households of slaves tended to be more stable, and the risk of arbitrary dissolution

was small though never totally absent. This was true of most of Latin America after the eighteenth century, the main exceptions being the mining areas and those regions in which the plantation system became dominant.”

The Murder of Slaves Because slavery is always a relationship that rests ultimately on force, it is hardly surprising that in every slave society the master has the power to inflict corporal punishment. As a matter of fact, throughout the precapitalist world and in a good many modern societies the paterfamilias has the power to physically castigate his wife, children, and servants. Recently an English court ruled that a man has the right to punish his wife by slapping her on the behind. The problem, then, is how far a master was entitled to go in his disciplinary actions. Two questions must be differentiated: the master’s right of life and death over his slaves (the jus vitae necisque) and his overall treatment

of his slaves. The two often vary together; that is, where masters could kill their slaves with impunity, they tended to treat them harshly. But this is only

a probability: there were many instances where masters could kill their slaves under special circumstances, or even whenever they pleased, yet in general treated their slaves relatively well. This was the case, for example, among the marsh Arabs of southern Iraq where a master was permitted to kill his female slave at will because he owned her “in blood and bone.” This power, however, was rarely exercised. On the contrary, the female slave was well treated, since more often than not she was her master’s concubine.” Societies varied considerably in the degree to which their legal codes or

The Condition of Slavery 191 customs permitted the murder of slaves by their masters. At one extreme were societies in which not only did the laws recognize such a power, but the masters frequently exercised it. Among the Goajiros of South America, for

example, masters could “at any time ... kill any of their slaves, man or woman.””> Nothing in the annals of slavery, however, can match the Indians of the U.S. northwest coast for the number of excuses a master had for killing his slaves and the sheer sadism with which he destroyed them. Among the Aleut slaves were killed simply to placate the grief of their masters when a son or nephew had died accidentally. At such times “they drowned them in water, threw them off a cliff in the sight of their parents, in whose despair

and bereavement they hoped to find their consolation.”’° Alternatively, slaves were killed to celebrate some special event such as a son’s becoming a shaman.’’ In almost all these tribes slaves were killed upon the death of their owners, especially when the latter were important persons: among the Tlin-

git, the selected slaves were bound hand and foot and thrown alive on the funeral pyre.’” When a new house was to be built, they were killed and buried beneath the posts;’”” at ceremonies of initiation, especially into the cannibal society of the Kwakiutl (according to Franz Boas), the body was torn into little pieces and eaten by the initiates.*° But it was during the potlatch ceremony culminating in the ritualized exchange and destruction of property that the murder of slaves became a veritable carnage, in which “rival leaders attempted to surpass one another in the number of slaves killed.””*’

The murder of slaves for ritual purposes was, of course, widespread. It existed, at some time, on every continent and in the early periods of every major civilization. Vast numbers of slaves were buried, often alive, with the earliest Chinese emperors.” In Japan, between the second and third centuries B.C., as many as a hundred slaves were buried with an empress.** The practice was widespread in the ancient Near East,** and among most early European and Asian peoples.*? We have a vivid account of one such gruesome ceremony of the Vikings from an Arab ambassador who lived among them in the early tenth century.*° The Aztec slaughter of thousands of prisoners of war and slaves bought for the purpose is well known,” as is the similar ritual slaughter by the Dahomeans**—although in both cases allowance

must be made for exaggeration and propagandistic bias in the original sources. The ritual murder of slaves, it should be noted, does not necessarily imply that masters had the capacity to kill slaves in other contexts. Among the Margi of Nigeria®’ and the Ashanti of Ghana,” who practiced human sacrifice, a master who indiscriminately put his slave to death could suffer the death penalty himself. At the other extreme were slaveholding societies in which the murder of slaves was punished in the same way as the murder of free persons. In some of the Southeast Asian states such as ancient Vietnam and Thailand, masters

192 Slavery as an Institutional Process who killed their slaves could be punished “according to law.”’”' Ancient Hebrew law was far superior to the other codes of the ancient Near East in this regard: “if the slave died on the same day that he was beaten by his master, the death was treated as murder.””* In medieval Europe the influence of the

church led to the enactment of penalties for the murder of slaves by their masters. Excommunication for two years was the penalty laid down by the Council of Epaone in 517, and Theodore, archbishop of Canterbury, prescribed a penance of seven years for a mistress who killed her slave in anger. It was rare, however, for masters to suffer the same penalty for killing one of their slaves as for killing a free person. Although the master’s right of life

and death was limited in thirteenth-century Spain, the penalty was still light.”’> And whatever the formal position taken in religious or civil law, it was extremely unusual to find a master in practice suffering the death penalty for the murder of his slave. Here and there we come across the excep-

tional case. In Siena one Giovanni de Sutri, a man-at-arms, was in 1436 “sentenced by the Podesta to have his head cut off, for having killed his own slave with a knife, ‘contra forman juris et statutorum Senensis.’ ’’ More typical was the case of Bartolomeo de Prata, who two years later was sentenced to pay a moderate fine for the same offense.”

Most slaveholding societies fall between these two extremes. Ancient Athens is typical. From the seventh century on, the murder of slaves was a legal offense, although the penalty was much less severe than for the murder of a free person. The law, however, was meaningless because the murdered slave’s kinsmen, invariably also slaves, had no way of bringing a case against the master. Glenn Morrow, who tried his best to interpret the data in a favorable light, was forced to conclude that “the murder of a slave could often escape without punishment.””° In Rome it was not until the first century A.D. that some restraint was placed on the power of the master to kill his slave, and this was only with regard to the practice of sending one’s slaves to fight with wild beasts. More meaningful curbs came with Antoninus in the middle of the second century.”°

The Greek and Roman experiences point to the major problem of the slave in the vast majority of slaveholding societies, from the most primitive to the most advanced: slaves were not allowed to be witnesses or to swear oaths, except under very special circumstances such as the trial of their master for treason. Even in these exceptional cases the slave’s evidence was usually taken under torture. In oriental and Western law slaves were allowed to take their masters to court only on matters relating to their own emancipation. A few post-Roman law codes added the capacity to go to court over disputes relating to the peculium, but actual cases were rare indeed. So too were cases involving contracts between masters and their slaves: one such case, described by Charles Verlinden as a “very curious” one in the history of the peculium, took place in Spain in 1284.’ The truth is

The Condition of Slavery 193 that in almost all slaveholding societies the master—in practice, if not in legal theory—had unlimited power, including that of life and death, over his

slave. The Roman jurist Gaius was right in declaring that the jus vitae necisque (right of life and death) belonged to the jus gentium.”*® To examine the legal capacity of masters to kill their slaves, I assembled data on forty-five of the societies in the sample of world cultures and coded

them by whether the penalty for killing a slave was (1) the same as that for the murder of a free person; (2) not the same, but very severe; (3) mild, amounting to no more than a small fine; or (4) negligible—the master was able to kill his slave with impunity. I found that nine societies (20 percent) fell into the first group; two (4.4 percent) into the second; six (13.3 percent) into the third; and twenty-eight (62.2 percent) into the fourth group. There are some interesting correlates to this variable. Where masters are allowed to kill their slaves with impunity, there is a greater tendency not to sanction the peculium; crimes against slaves are punished less severely than crimes against nonslaves; masters rely more heavily on captivity and kidnapping as the means of enslavement; and there is less tendency to rely on birth and punishment for crimes as sources. Race, ethnicity, and dominant subsistence patterns were found to have no influence on the legal capacity of the master to kill his slave. One rather odd relationship stands out. The legal capacity to kill one’s slave is closely associated with a low proportion of males in the servile population. Since there is an equally surprising relationship between this variable and others measuring the treatment of slaves, it is best to look at all the relevant factors before attempting to explain it.

Crimes against Slaves by Third Parties So far we have considered only the treatment of the slave by his master. What about the delicts of third parties against slaves? Here laws were likely to be more important in influencing practice, for the master had a vested interest in the protection of his slave property. Proprietary self-interest, how-

ever, benefited the slave in very limited ways. There were many crimes against the person of the slave for which the master did not consider it worth his while to seek redress, for example, the rape of a female slave which resulted in no damage to her working capacity, or the mortifying verbal abuse or slap in the face of a male slave. In no slaveholding society was the honor or dignity of the slave taken into account in law or practice. The comparative data show, further, that in most slaveholding societies delicts against slaves were penalized not in the interest of the slave but in order to protect the master’s property. Murder was the one major exception, for most societies considered the killing of a slave by a third party not only an assault on another man’s property but sinful. Even with murder, it was usually the case

194 Slavery as an Institutional Process that punishment beyond the payment of damages was rarely very severe. Sometimes the emphasis on compensation to the master for the injury of the slave by a third party had bizarre and tragic consequences for the slave. Among the Bedouins of the Sinai, damage to a slave short of murder was compensated in the same way as damage to a woman, and this could be very expensive. For the murder of a slave, however, there was a fixed payment:

two camels. Hence it was common to hear members of the Towara tribe saying that if you injured a slave “it was cheaper to kill him outright.””’ Rarely did the slave receive any compensation for his injuries; when he

did it was a mere pittance, almost an insult. In Iceland, for example, as among all the Germanic peoples, no wergild was paid to the kinsmen of the thrall when he was murdered, only to the master; and if the thrall received a heavy blow from someone other than his master, he was given a compensa-

tion of three aurar, equivalent to three cows (an average male slave was worth twelve aurar).!™

The Greek view of the matter was typical of many precapitalist (especially tribal) peoples. The murder of a slave was considered a form of ritual pollution requiring religious purification for the good of the community rather than for any consideration of the injustice done the slave,'°' a view found in several tribal slaveholding groups.'°* On Greece, Glenn Morrow writes: “That it was something more than a punitive fine, something less than death—this is about all we can infer as to the penalty for murdering a slave.”!?°

A slave could no more give evidence against a free third party than he could against his master.'°* Only the master could take legal action against the third party for injuring or murdering his slave, and in actuality it was rare for him to do so. Slaves could lodge information with a magistrate on which prosecution against a free person, his master or anyone else, could be based, but it is not clear how much good this did the slaves.'”” The information would be useless unless a free person was prepared to stand witness on behalf of the injured slave. It is improbable that any Athenian would testify against another free citizen; in the unlikely event that such a person existed, it would obviate the need for the slave to lodge the information in the first place. Clearly the provision was meant for slaves to act as spies against their masters and other free persons who had committed serious crimes against the state. It was not, and could not have been meant, for the benefit of the Slaves as Morrow implies. The Russian classicist E. Grace has shown that Athenian homicide laws

respecting slavery emerged slowly as part of the development of laws discriminating between citizens and noncitizens. She speculates that the legal autonomy of the master grew with the emergence of large-scale slavery.'”° She also argues that slaves were not treated as a homogeneous group. “How he would be dealt with as a killer, the extent and type of intervention by

The Condition of Slavery 195 public authority in prosecuting his punishment, might well depend upon whose slave he was, a citizen’s, a metic’s or an altogether foreigner’s.”'°” His

treatment as a criminal would also depend on the status of his victim, “‘in particular whether or not he was a citizen. The shedding of a citizen’s blood, even perhaps the blood of a citizen’s slaves, could have different legal consequences from those obtaining when the victim too was an outsider.”'? We know almost nothing about the nature or relative degree of punishment meted out to slaves, as opposed to nonslaves. It does appear that slaves were subject to torture more often than other groups, and such torture was as much a form of punishment as it was a means of gaining information.'” The principle of privilege was an entrenched feature of Roman law (as we have already seen). Until the early empire, the victimized slave could seek redress only through his master. The rule that a slave could not give evidence against a free person except for special cases in the public interest (and then, under torture) was progressively modified under the empire. The murder of slaves by freemen without cause was punished first by banishment and later by execution.''® For lesser injuries against the slave, compensation had to be made to the owner, and from very early we find operating in Roman law the near universal principle that crimes against slaves were punished less severely than comparable crimes against freemen. The Laws of the Twelve Tables, for instance, laid down that the punishment for breaking a slave’s bones was half the fine due when the victim was a freeman.''' Criminal slaves were either handed over to their victims or punished by their masters, but a senatorial decree of A.D. 20 put a stop to this: slaves from then on were tried in the same manner as the lowest-ranked free persons.''” As in Greece, distinctions were made among slaves and their victims in the treatment of both the crimes and the injuries of slaves. The slaves of powerful patricians were clearly out of bounds for proletarian bullies. We shall see in Chapter 11 that the familia Caesaris, itself a heterogeneous group, had privileges other slaves did not. This sort of status distinction in the treatment of slaves is found in nearly all premodern and modern slave societies. Among lineage-based societies, for example, the royal slaves of the Igala kingdom of West Africa, especially the amonoji (palace eunuchs who protected the person and ritual space sur-

rounding the king), held a privileged status vis-a-vis other slaves.''? And among the Somali, nothing short of blood payment was demanded if the killer of a man’s slave was a member of the outcaste Sab, whereas only money compensation was demanded if the killer or injurer was a fellow Somali freeman.''* In the pre-European Islamic states of Malaysia the murder of an ordinary slave by a freeman usually went unpunished except for payment of the slave’s market price, whereas the murder (or even the injury) of one of the raja’s slaves was punishable by death. Ethnic factors also played a part. Members of the Arab ruling class were

196 Slavery as an Institutional Process never punished for the murder or injury of a slave, whatever the latter’s status, only members of the indigenous Malay community.''’ More privileged

too were the government and church-owned slaves of Visigothic Spain. Skilled household slaves there, the idonei, could expect better treatment from free third parties than the rustici, or rural slaves.''® And in thirteenthcentury Spain, with its bewildering variety of races, ethnic groups, and creeds, the condition and treatment of slaves both as victims and as criminals varied widely. Orthodox Greek slaves, for example, had a special status

in law compared with other slaves, and Jewish and Muslim masters had greatly delimited powers over their slaves compared to their Christian counterparts. |"

The Delicts of Slaves No slave society took the position that the slave, being a thing, could not be held responsible for his actions. On the contrary, the slave usually paid more heavily for his crimes when the victim was a freeman. It was different, however, when slaves committed crimes against one another. The penalties then were usually much lower than those for crimes among freemen or involving freemen and slaves. Islamic law was typical in holding that the falio (an eye for an eye) did not apply to slaves. The reason had little to do with any con-

cept of reduced responsibility; rather, it was out of consideration for the owner’s interests. If a master had already lost one of his slaves at the hands

of another, he was hardly inclined in the name of some abstract sense of justice to deprive himself of the services and value of the offender. A beating was about as much punishment as he was prepared to inflict if the determination was left to him (as it usually was, with respect to crimes committed among slaves of the same master).''® Where the offense involved slaves belonging to different masters, the tendency almost always was for the masters to settle the matter between themselves, unless there were additional grievances. Roman imperial law was unusual in having the courts take over the punishment of slaves for serious crimes against other slaves, and it was noteworthy that this development came relatively late. Although in theory the slave had no will beyond that of his master, in no slave society was the master held responsible for criminal actions, especially

murder, committed by his slaves against free third parties—unless, of course, he ordered the crimes. The criminal slave was usually handed over to the relevant state authorities, or, where there were no formal legal organizations, he was delivered to the victim’s kinsmen. Treatment in Iceland was typical: “The thrall was held responsible by the law for the deeds of violence he was accused of having committed. The placing of this responsibility on him is the best evidence that he was considered to be a human being relative to his crime ... It was not due to any benevolence on the part of the masters

The Condition of Slavery 197 toward the thrall that he was considered to be a human being when he was in the toils of the law. This law is a punitive measure, promulgated for the purpose of safeguarding the rule of the master class,” and only in exceptional cases was he given an opportunity to prove his innocence.''” The position of alienation of responsibility adopted in Visigothic Spain, derived from Roman and Germanic laws, is true of all slave societies in their response to the master’s responsibility for crimes committed by his slaves against free third parties.'“” Even among the Ashanti, with their extreme form of the principle that the head of the household 1s totally responsible for the actions of its members, an exception was made for the murder of third parties by slaves. “A master is not killed for the murder of his slaves,” goes the validating legal proverb.'*’ The Fanti, another Akan group, echo the same principle in their maxim, “A slave does not commit murder for the master.” !7”

The issue becomes somewhat more complex with respect to the torts of slaves involving third parties, or lesser crimes punishable by fines. The slave was expected to pay the fine or compensation from his peculium, but if the peculium could not cover it, the master often had to pay, then confiscate the slave’s peculium and make up the balance in other, no doubt punitive, ways. Far more complex were cases involving commercial transactions between slaves and free persons. If it was clearly understood that the slave was acting with his master’s permission, but on his own behalf, then only his peculium was liable for forfeiture. The law courts of all commercially advanced slaveholding societies permitted slaves to engage in such transactions, Rome being the classic case.'** The master, of course, was the person sued, since the slave was not allowed to engage in contracts, to be sued or to sue, but apart from unusual cases the master could only be sued “dumtaxat e peculio, up to but not in excess of the value of the slave’s peculium.”'** However, in slaveholding societies that had any measure of commercial activity, trusted and skilled slaves acted as agents on their master’s behalf. Where the matter was clear-cut and the slave understood to be an agent, there were few problems: the master was obviously liable and the slave’s peculium did not enter the picture. But the issue was often anything but clear-cut. The peculium of

the slave might have been consolidated with other areas of the master’s property and thorny legal problems created if and when liability had to be assessed. A vast section of Roman commercial law was devoted to this problem. The general rules that evolved were, first, that the slave, being a person without authority and powerless, “could not transfer dominium. If he sold and delivered possession passed but no more.” Second, “a slave {had]

no authority to make his master’s position worse.”'”° It is easy to see the kind of opportunities these rules offered to unscrupulous owners.

198 Slavery as an Institutional Process

The Overall Treatment of Slaves In view of the many factors that influenced slavery in any given society, some means of assessing the overall treatment of slaves was highly desirable. I therefore developed a four-point scale to code those societies in the world

sample on which reliable data were available. Fifty-eight societies were coded on the following points: (1) whether there were some legal restraints on masters and third parties in their relations with slaves, and in practice slaves were treated well; (2) whether there were no formal restraints, but in practice slaves were treated well; (3) whether there were some formal restraints, but in practice slaves were treated badly; and (4) whether there were no formal restraints, and in practice slaves were treated badly or brutally. The following frequency distribution emerged:

, , No legal oo No legal

, Legal restraints; —_— restraints; Legal restraints; restraints;

Number 29 17 3 9 total SO 29 5.5 15.5

Societies treatment mild treatment mild treatment harsh _ treatment harsh Percent of

the societies. | ,

We see that the treatment of slaves was judged mild in nearly 80 percent of The relative size of the slave population does not in itself significantly influence the treatment of slaves. Where there are few slaves, there is a greater tendency for them to be integrated into the households of their masters and

into the traditional work patterns alongside nonslave persons. Where the number of slaves grows larger, the rest of the population is obliged to take a more formal interest in them—if not for the slave’s well-being, at least to prevent the expression of antisocial and wantonly cruel behavior, which may spill over into the relations between free persons and offend communal norms. As the slave population grows, however, there is an increasing tendency to segregate it, usually in specialized productive activities. This development cuts both ways as far as treatment of the slaves is concerned. Where

mining, latifundia, or plantation-type farming systems prevail, the slave tends to become merely a unit of production. The typical slave is in a large holding and has little personal contact with his master. Overseers, themselves frequently slaves, control him. The whip becomes the major impetus to work, and brutality increases. The classic cases here are the non-Latin Caribbean slave societies, nineteenth-century Cuba, and the plantation belt in Brazil; and with respect to mining, the Laurium mines of ancient Athens and many though not all those of the Americas. In many societies with large

The Condition of Slavery 199 slave populations the slaves are used as tenant farmers. While there are often “supervisors,” their role and functions tend to be different from those of the slave overseers of the Americas. Slaves have a high level of social and economic independence in such systems. The best examples are the slave systems of the Sokoto caliphate during the nineteenth century. Ethnicity was found to bear no relation whatever to the treatment of slaves, and the fact that slaves were of a different race from their masters did not significantly influence their treatment, even where masters had strongly racist views. As we indicated earlier, it was surprising to find that societies in which female slaves outnumbered males were those in which masters were most

likely to be able to kill their slaves with impunity. It was found too that crimes against slaves were less punished in societies where female slaves outnumbered males. At the same time it was found that the sex ratio of the slave population was not significantly related to the overall treatment of slaves. How do we explain these relationships? It is important to note, first, that female slaves outnumber males in 54 percent of all the slaveholding societies in the Murdock sample; they are equal in number to men in I7 percent; and number less than males in only 29 percent of the sampled societies. Yet the sex ratio of the slave population bears no relation whatever to the proportion of the population enslaved, nor to racial and ethnic differences, nor to the mode of organization of the slave population. These negative findings are consistent with the absence of any significant relationship between the sex ratio of the slave population and the overall treatment of slaves. The clue to the peculiar set of relationships between the sex ratio and the more specific variables is the fact that societies with more female slaves tended, to a greater degree than those with more males, to be ones in which household production prevailed. In such societies the master, as patria potestas, usually had the power to discipline to the point of death all members of the household, not only slaves but wives, children, junior kinsmen, and retainers. We are dealing here with typical kin-based slaveholding societies such as those described in Homer, primitive and early republican Rome, many traditional African societies, and many of the less advanced Islamic and oriental societies. At the same time it is in precisely these societies that the female slave was very rapidly assimilated to the status of junior wife or fictive kinsman. She may have been killed with impunity, for she belonged in “blood and bone,” but under the master’s potestas this happened no more frequently than it did to “free” persons.

| The Slave as an Active Agent So far the slave has figured in our discussion as a passive creature largely at the mercy of forces beyond his control. Within the law of all slaveholding

200 Slavery as an Institutional Process societies, this was pretty much how he was conceived. The slave, of course, could always act outside the framework in which he was legally or even socially defined. In taking account of him as a criminal, I have already indicated that in no slave society did the slave accept his lot or his legal definition. In another work I have considered the slave as rebel.'”° It is significant that the slave as an active agent was recognized only when he behaved in a

criminal manner. Are there any exceptions to this general rule? Was the slave’s capacity to act on his own behalf ever recognized in law, apart from the very special cases discussed earlier? The answer is yes. There were three actions of a positive, willful nature that many slaveholding societies recognized. These were the capacities of the slave to defend himself from the murderous assaults of free third parties; to seek sanctuary; and, in extreme cases, to change masters. SELF-DEFENSE

While the slave was rarely allowed to seek redress in court for injuries done him by third parties, a significant minority of slave societies allowed him to defend himself from unprovoked attack and even, in some cases, to defend his woman. Quite apart from any appeal to simple justice, such defenses were legally implied in the universal existence of laws that forbade others to damage a free man’s property: in defending himself, the slave was defending his master’s interests. In medieval Iceland a thrall was permitted to kill a freeman who attempted to violate his “bedfellow” or to harm his own person. However, the thrall stood a very slim chance of ever committing such

an act of defense; every free Icelander went about armed to the teeth, whereas thralls were rarely allowed to carry arms of any sort. In Norway this

“right” was somewhat backhandedly given: the thrall was permitted to throw a bucket of water over the couple who had offended him. As Peter Foote and David M. Wilson comment, “The contempt for the seducer here only matches the contempt for the thrall.”’!?’ Slaves were allowed to defend themselves and their women in traditional Malay law, especially in Jahore, “for it 1s written that no married woman shall be made light of; this is the

law of custom, but by the law of God whoever kills shall himself be killed.”'** The female slave could also be protected from sexual abuse by her spouse in China, Korea, Vietnam, and among the Ashanti, but in all these cases the slave faced the same problem as his Icelandic counterpart: he

was usually unarmed, and he had no way of defending himself in court against a later charge of injury by the freeman in the event that the latter had not been killed. Where it existed, this was a privilege which none but the most daring slave would choose to exercise. In any event it was a privilege found in only a minority of slaveholding societies. Most forbade the slave to defend himself by force except at the order of his own master.

The Condition of Slavery 201 SANCTUARY

The capacities to seek sanctuary and to change masters were closely related,

in that the granting of sanctuary was sometimes followed by a change of master and, alternatively, the slave’s plea to another master to buy him often amounted to a plea for sanctuary. For purposes of clarity, however, the two will be considered separately here. Sanctuary was most frequently granted by religious organizations. The Ashanti and the Nyinba, a Tibetan-speaking Nepalese group, were typical. Among the Ashanti “a slave could run away and seek sanctuary by throwing himself on the mercy of a god (e.g., Dente) or the ‘samonfo’ (ancestral spirit)

at some Barim (mausoleum)”; and among the Nyinba a harrassed slave sought sanctuary in the temple of the local deity of a neighboring village.'””

The practice was widespread in ancient Greece: in Gortyna, it seems the slave in sanctuary could be protected even against the force of law; in Athens, it was the only procedure a slave could invoke on his behalf. The custom was so entrenched in Hellenistic Greece that it continued unabated during Roman times, much to the annoyance of Roman administrators. A Roman praetor once was forcibly prevented from removing a slave from the

shrine of Diana at Ephesus and Tacitus complained of the abuse of sanctuary in Asia.'*° But the Romans themselves, by the time of Tiberius and possibly earlier, came to allow the slave to seek sanctuary not only in the old religious manner, but at statues of the emperor as well. While the Christian church took over the pagan custom of offering sanctuary to slaves, it never had the same power or desire to protect the slave in sanctuary.'*! An incident recorded by Gregory of Tours reveals the ineffective high-mindedness that characterized the church, not only during the late empire but throughout the Middle Ages. Two slaves who had been forbidden to wed fled to the church in order to get married and to seek sanctuary.

The master demanded them back, indicating that he intended to punish them. “You cannot receive them back,” said the priest, “unless you pledge that their union shall be permanent, and that they remain free from all corporal punishment.” The master promised, the slaves were released, and as soon as they were back in their master’s clutches they were severely punished.'* The only time the church really went out of its way to help the slave was when the latter had absconded from a Jewish master. And, as Iris Origo observes, this was wholly because of religious bigotry, for slaves were on the whole better treated by Jewish masters.'*” In some societies lay officials and even ordinary freemen could offer sanctuary. The statue of the emperor was in fact a secular form of sanctuary, although it had overtones of emperor worship. In the pre-European Malay states a slave in trouble could become a hostage of the raja, a practice the rajas did not discourage, since the slaves thereby became their own.'** There were also cases of influential lay individuals offering sanctuary.

202 Slavery as an Institutional Process Among the Sherbro of Sierra Leone, a slave could seek the intervention of an influential third party who would “sit near” him as his protector. To regain his slave, the master had to get a group of officials to “beg” the slave to return. The officials would then accompany the slave to his master, but would only turn him over after the master had publicly spoken well of the slave and indicated his sincerity by offering small gifts to the officials. A fascinating aspect of this practice is that it was the same procedure used to regain a disgruntled wife who had fled to the sanctuary of her own family.'*’ According to Lombard law, any free man could offer sanctuary to a slave.'** This may have worked in tribal times, but from the severe penalties inflicted on free persons harboring runaways in Tuscany during the fourteenth and fifteenth centuries, it is clear that only the most foolhardy or powerful Lombard lord would perform such a function.'*’ CHANGE OF MASTERS

Finally, we come to the privilege of changing masters. The Ashanti sum up the normal attitude of most slaveholding peoples in one of their bluntest proverbs: “A slave does not choose his master.” Yet this was not always the case. Even among the Ashanti, a slave who felt extremely badly used could, as a last resort, swear an oath that some other master must buy him.'** The capacity to change masters was found in several advanced, large-scale slave

societies—although the degree to which law was reflected in practice is questionable. By the middle of the second century A.D. cruelly used slaves in Rome could not only seek sanctuary at the statue of the emperor but, in the-

ory, request too that they be sold to another master. Spanish law and its Latin American variant also required that a cruelly treated slave be sold to another master. Still, it was unusual for such action to be taken on the slave’s behalf in either Spain or Latin America. The Colombian Choco during the eighteenth century was one such exception. Even in Colombia the courts took this action only in cases of excessive brutality. Furthermore, the slave had no say in the choice of the master to whom he was sold. The action was more a mild punishment of the master than a “right” of the slave.'”” The ability to change masters was surprisingly widespread because of a peculiar Islamic custom that permitted it. The custom, sometimes in odd forms, was found all over Muslim Africa, the Middle East, and Islamic Asia. The most common version was for the disgruntled slave to go to the compound of the master he wished to have buy him and cut a piece of flesh from the ear of that master’s camel or horse. The owner of the slave was then required to compensate for the damage by handing the slave over to the offended master.'*° A more bizarre variant found among the Wolof and the Sereer involved cutting off either the ear of the intended master’s horse or the ear of the intended master himself!'*' In parts of Saudi Arabia it was the

The Condition of Slavery 203 property of the present, not the intended, master that was mutilated. According to an Arab custom known as beja, the owner of a house was responsible for any delict committed in front of it. Disgruntled slaves took advantage of this by bringing the camel of their present master to the door of the intended master’s household and slaying it there. “Usually the man in front of whose house the slaying occurs [preferred] to repay one camel and gain a

tom.”!** |

slave, rather than kill the slave and suffer dishonor from a broken cus-

Although this curious practice was widespread throughout the Islamic world, it is extraordinary that its origins are unknown. Koranic law enjoins the master to provide for the welfare of his slave or else sell him to another, but there is nothing in the Koran or early religious traditions that would account for this strange mode of changing the master. The French anthropologist André Bourgeot, who has worked among the Imuhag group of the Tuaregs, offers a symbolic and historical explanation. He argues that the custom is a symbolic reaction against the mutilation of slaves during the pagan days of the Arabs, a practice soon condemned by Islam as the inspiration of the devil. Hence, cutting the ear of the camel symbolizes bad treatment and un-Islamic behavior, by harking back to the pagan practice of slave mutilation. The former master for this reason is not permitted to intervene, and the flight of his slave represents both a social and an economic loss. The act, Bourgeot further claims, is an incipient form of individual rebellion on the part of the slave, though one couched in religious rather than political terms.'* This interpretation is attractive and may well hold for the Tuareg and other Islamic peoples. The problem is that several pagan or other nonIslamic peoples have a closely related practice. An almost identical custom

is found among the Toradja-speaking pagans of the central Celebes. N. Adriani and A. C. Kruyt, who lived among these peoples at the turn of the century, found that a slave who was wronged and wanted to change his master went to the residence of another lord and either broke up some of his furniture or burned pieces of his clothing. The practice was called mepone, which literally means to climb up to another’s house. When the master came for his slave he had to pay a buffalo for him, and this was meant to teach the master a lesson. But if the slave really wanted to have his master changed, he

would destroy so much of the intended master’s property that the present master had no choice but to turn him over to the offended master as compensation. Another method involved cutting a lock of hair from a member of the intended master’s household and then burning it.'** Although some groups of the Toradja were subject to Islamic influence, the pagan tribes studied by these authors were not; so we may rule out any diffusionary explanation. Furthermore, we find a basically similar practice halfway around the world from the Toradja. Among the pagan Ila of Central Africa, a dis-

204 Slavery as an Institutional Process satisfied slave who could not find another master to buy him would go to a neighboring village and insult the head of the kin group by throwing ashes on him. The fine for this outrage was two beasts, but the slave was usually accepted as a compensation if that was his intention.'* The comparative data, then, offer no support for Bougeot’s symbolist explanation. The practice, I would argue, is better explained in terms of the general condition of slavery itself. The slave everywhere was held as an extension of his master, a vocal instrument without will except where this worked against the slave. By performing this act of mutilation, the slave achieved three objectives in one fell swoop. First, he succeeded in changing his master. Second, he gained revenge against the master, in that the latter always came out a net loser. (It is significant that when the damage was to the intended master’s property, it was only superficial; but when, as in Saudi Arabia, the damage was done to property belonging to the present master, it was real and substantial—a camel was slain, no doubt the master’s favorite.) Third, the slave, if only temporarily, asserted his will. This was indeed an act of individual rebellion—not political, to be sure, but not religious either. Rather, it belonged to the category that Albert Camus calls existential. It is not by accident that Camus opens his great work, The Rebel, by citing the slave rebel as the archetypal existential rebel. By saying no, the slave set limits beyond which he could not and would not be demeaned. He demanded a recognition of his humanity not only in a manner that worked against him—tfor every slave came to realize existentially what Camus arrived at intellectually: that it is not possible to deny another person his humanity, the worst the oppressor can do is recognize it negatively and exploitatively—but in a manner that worked for him, in his own interest. This was the true interpretation of the practice. Seen in this light, it is not a peculiar custom of Islamic culture but an imperative of the slave condition. The fact cannot be gainsaid, however, that exchange of masters was found overwhelmingly in Islamic lands. Why 1s this? The answer is that Islamic law enjoins the master to change his slave if the latter is excessively unhappy with his treatment. The normal method of changing masters was simply for the slave to demand such a change, which the master was duty bound to honor. The custom of abusing the proposed master’s property was used only when the master refused to obey the dictates of his religion. Thus, far from being a symbolic harking back to pagan times, the custom was an affirmation of Islamic religious law. This law, however, only made possible the expression of the practice; indeed, the custom was obviated if the law was followed. Islamic law does not explain the custom as such; it only explains the frequency of its occurrence in Islamic societies. The practice itself is explained by factors inherent in the condition of slavery. At the same time, the absence of any requirement in the other major religions to sell an unhappy slave explains the infrequency of the practice else-

The Condition of Slavery 205 where. This does not mean that the imperatives of slavery were not present in such societies, or that the slave’s yearning for vengeance and the positive affirmation of his will were not as pressing. It is, rather, that they had to be expressed in other ways. The ground rules laid down by the master class determined how the slave reacted, how he manipulated or, when necessary, broke the rules. The fact that an extremely unhappy slave could have his master changed was an important safety valve. Other slaveholding peoples have had other safety valves—liberal laws on the peculium, a high rate of manumission, a sufficient incidence of slaves allowed to hire themselves out and live apart from their masters, effective forms of sanctuary, fictive kinship assimilation, adequate physical treatment, to list some of the more familiar. And still other slaveholding societies had no safety valves at all, relying mainly on brute force.

Conclusions The limited nature of the data on most premodern slaveholding societies obliges us to rely not only on gross indexes of the condition of slavery but also (perhaps too heavily) on legal norms. For most slaveholding societies we simply do not know how often slaves were beaten and for what reasons, much less the internal factors that determined the frequency of the use of violence. Nor can we compare treatment in terms of such “welfare” variables as diet, clothing, housing, and health. And of course we know almost nothing directly about the way slaves felt about their condition in the great majority of such societies. With all their limitations, however, the data nonetheless allow us to infer certain general statements about the condition of slavery. The most important conclusion is that the master-slave relationship was not a static one in which an active master constantly got his way against a wholly passive slave. In spite of the extreme power of the master, certain constraints were inherent in the very nature of this relationship. One was the self-interest of the master himself. The whole point of keeping slaves was to get them to serve him, in whatever capacity he chose, to the best of their ability. To achieve

this objective the master could use various combinations of punishments and rewards. Slavery was unusual in the extraordinary extent to which the slave could be punished for not serving—even to the extreme of murder. But a dead slave, or one incapacitated by brutalization, was a useless slave. This stark fact, plus the recognition that incentives usually work more effectively than punishment in inducing service, was enough to encourage most masters in all slaveholding societies to search for the best balance between reward and punishment. The second category of constraints was the slaves themselves. Powerless, isolated, and degraded in the eyes of nonslaves they may have been, but they

206 ‘Slavery as an Institutional Process struggled constantly to set limits beyond which they would not be expected to go. In so doing, they could regularize their relationships with the persons who. parasitized them and carve out some measure of predictability, if never legitimacy, in their social behavior. Both masters and slaves, then, had to adjust to one another. Just how much the master would concede in order to gain, and just how much of the master’s parasitism the slave would take before declaring a limit, varied considerably both within slaveholding societies and between them. It is trite— and untrue—to say that the relationship that emerged varied with the char-

acter of the master. For even on an individual level the condition of the slave varied with the character of the slave also, as well as with the social and economic circumstances that impinged on the dynamics of the interaction between slaveholder and slave. The experience of the U.S. South, which presents us with a richer body of data than any other slaveholding society (including the testimony of hundreds of ex-slaves), fully supports this general conclusion. From his study of

the W.P.A. and Fisk University slave narrative collections, Stephen C. Crawford found that in spite of the total legal power of the masters, self-interest and the determination of the slave to survive as best he or she could created an environment in which slaves could even “significantly control

their personal probability of punishment.”!*° Self-interest dictated that punishment was not used mainly as a labor promoter but as a form of social

control. Slaves also had some option in the kind of household structures they established, and appropriate choice of such structures could reduce the risk of punishment of children and sale away from family and friends.'*’ None of this, of course, implies that the system was not oppressively weighted against the slave. Beyond the character of the master, there were

other key features in the environment that the slaves could do nothing about. One was the size of the farm unit on which they lived, another was its

location. Both were crucial in determining the condition of slaves in the United States,'** as they were in all other advanced slave systems. Large farms meant a higher level of whippings, less contact with owners, fewer chances therefore to manipulate the political psychology of the relation, and more work. But even if the slave had some choice in the size and location of the farm on which he lived, he would still be faced with a no-win situation. Small farms, while physically less demanding, offering more opportunities to acquire skills, and allowing far more contact with (and manipulation of)

the owner, had their own special horrors. More personal contact meant greater exposure to sexual exploitation for slave women, including the not infrequent experience of gang rape by adolescent kinsmen of the owner. The probabilities of family breakup as the result of such sexual exploitation, and of being sold away, were also greater on such farms.'*’ These examples give some idea of the complexities of both the relation-

, The Condition of Slavery 207 ship itself and the conditions that influenced it. In macrosociological terms the U.S. South, as I have repeatedly emphasized (and will have reason to do again), was rather unusual. But it shared with all slaveholding societies certain imperatives of the interaction between slaveholder and slave. Among these was the fact that in southern slavery, as in all other slaveries, there was a constant struggle between master and slave in the effort of the former to gain as much as possible for himself with the least possible loss, including the self-defeating loss of his slave, and the effort of the latter to minimize the burden of his exploitation and enhance the regularity and predictability of his existence. It is a mistake to characterize such a highly asymmetric interaction as one of “give-and-take,” as Crawford does in his otherwise impressive study.'°” Husbands and wives give and take, sometimes; employers

and wage earners, maybe; masters and slaves, never. What masters and slaves do is struggle: sometimes noisily, more often quietly; sometimes violently, more often surreptitiously; infrequently with arms, always with the weapons of the mind and soul. In this conflict, as we have seen in Chapter 3, there was resentment on both sides—expressed in ideological stereotypes by both sides. But from

time to time masters who wrote about slaves and their condition broke through the barriers of their own prejudices and saw the relationship as it was, not only for themselves but for their slaves. Even in ancient India, where the ideological mystification of exploitation was taken to greater lengths than in almost any other society, masters nonetheless realized that slaves worked under duress and that behind all the rhetoric and religious reinforcement, naked force was the ultimate and essential sanction. During the Buddhist period masters realized that slaves worked resentfully even in the performance of religious work, despite the fact that all such work was considered meritorious when done by anyone, slave or free. In a celebrated passage from Majjhima Nikaya the slaves who were ordered to deputize for their masters in a very sacred ritual were seen as resenting the compulsion and “with tears in their faces, weeping, they [went] about their jobs.”!°' Nor were the masters of the Buddhist period always deceived by either their own

version of the “Sambo” ideology or by the aggressive duplicity of their slaves, for as another master observed: ““O Bhante, our slaves... do another

thing with their bodies, say another with their speech and have a third in their mind.”'°? The master then went on to explain what he meant by this, and in doing so went straight to the heart of the social and psychological struggle inherent in the relationship: On seeing the master, they rise up, take things from his hands, discarding this and taking that; others show a seat, fan him with a hand fan, wash his feet, thus doing all that needs be done. But in his absence, they do not even look if oil is being spilled, they do not turn to look even if there were a loss of

208 Slavery as an Institutional Process hundreds or thousands to the master. (This is how they behave differently with the body.) . .. Those who in the master’s presence praise him by saying, “Our master, our lord,” say all that is unutterable, all that they feel like saying once he is away. (This is how they behave differently in speech.)'”’

Masters and mistresses who saw beyond their ideological camouflage came to the same conclusion everywhere. The women of the master class usually saw what was going on better than their menfolk, because they had more time and leisure to reflect on the slave relation; but perhaps also because when they did reflect, their own condition as women in a timocratic culture gave them a better understanding of the struggle that underlay the apparent give-and-take of surface realities. It is no wonder that some of the best accounts of the interpersonal aspects of slavery in the antebellum South come from the diaries of women. Mary Boykin Chesnut, writing in the midst of the Civil War, tells of her first recognition of the reality beneath the surface: “They go about in their black masks, not a ripple of an emotion showing; and yet on all other subjects except the War they are the most excitable of all races. Now Dick might be a very respectable Egyptian Sphynx, so inscrutably silent is he. He did begin to inquire about General Richard Anderson. ‘He was my young Master once. I always will like him better than anybody else.’ ””!”*

Deep down, Chesnut undoubtedly knew what lay behind the mask, but it was too much for her to spell out, even in her diary. It is to the poet son of two runaway slaves, Paul Lawrence Dunbar, that we must turn for the answer:

We wear the mask that grins and lies, It hides our cheeks and shades our eyes, This debt we pay to human guile; With torn and bleeding hearts we smile.'””

|.e

Manumission: ~ Its Meani and Modes

ts Meanings

WE TURN NOw to the circumstances of release from the condition of slavery. The transition from slave to freed status posed many critical problems for a slaveholding society. What did manumission mean?! By what means did the transformation take place? How was the freedman incorporated into the society? In addition to these more cultural problems, social and statistical issues are involved. What were the conditions favoring manumission both within and between societies? Why were some slaves manumitted while others in a given society were not? And why did some slaveholding societies show a much higher rate of manumission than others? These questions will be explored in the next several chapters.

, | The Meaning of Manumission What was manumission, and how was it to be achieved? Unsuspected problems of extraordinary complexity arise when one tries to answer these apparently simple questions. Given the way in which slaves were legally and socioeconomically defined, there was no obvious way in which they could have been released from their condition. I call this the inalienability problem, and it is in examining the various ways in which the same or different peoples came to terms with it that we arrive at an understanding of the true

| 209

210 ~~ Slavery as an Institutional Process

nature of manumission. The problem may, for purposes of analysis, be divided into three parts: there is a conceptual issue; there is a cultural problem, which encompasses symbolic factors as well as legal or customary modes; and there is a social problem, which focuses on the status of freedmen. The first two issues will be considered in this chapter and the problem of status in the next. THE CONCEPTUAL PROBLEM

Common sense has it that the slave or someone else simply buys his freedom from the master. Such a view, itself narrow and problematic, is true only of the advanced capitalistic slave systems of the modern world. Almost no precapitalist slave system—including imperial Rome—considered the matter as

simple as this. Precisely because Rome had an advanced legal system but was also precapitalist, it highlights the problem in many striking ways. It was not altogether clear to Roman jurists that the sale of freedom was possible. Two reasons can be found for this view, and they apply as much to other precapitalist cultures as to ancient Rome. First, if everything the slave is and enjoys belongs to the master, then by definition it is never possible for the slave to buy back his freedom from his own resources. The peculium, it may be recalled, always belongs ultimately to the master, who allows his slave to enjoy its usufruct as long as he remains

a slave. It is not possible for a third party, either, to buy back the slave’s freedom, for that third party must either pay for the slave with money passed to him by the slave, in which case the master is cheated, or else pay for the slave with his own funds, in which case the problem of alienating slave from master is not solved but simply passed along to the third party, who now owns the slave. There is a second and more profound problem, namely, that it is impossible to express the idea of manumission in terms of any appropriate legaleconomic category. The most obvious legal institution is conveyance, but as

W. W. Buckland and others have shown, the manumission transaction is only analogous to conveyance, it is not identical to it.* In conveyance there is a seller with something to sell and a buyer who wants that particular thing. The buyer exchanges something else, say money, for the thing desired. The

seller, on acquiring the money, transmits the thing to the buyer. What the seller hands over and what the buyer receives are one and the same thing. Clearly, this is not what happens in a manumission transaction, even where the slave or someone else pays dearly for the slave’s release. For the master does not convey dominium or power to the slave; he merely releases him from his dominium. As Buckland puts it, “What passes to the man is not what belonged to the master, his liberty and civitas are not subtractions from

Manumission: Its Meaning and Modes 211 those of the dominus,” hence “what is released is something other than what 1S acquired.””° There surely has been a transaction, but whatever it 1s, it most certainly is not a sale as understood in Roman and all other legal codes. What is it then? Buckland proposes that the real meaning of manumis-

sion is that “it is not a transfer of dominium; it is a creation of a civis.’” Buckland is on the right track but leaves many questions unresolved. He is partly right in seeing manumission as a creation—but only partly so, for it is much more than that. Furthermore, while the thing created might have been a civis in the special case of Rome, this was not the case in most slaveholding societies, so the second half of his solution is of little value to someone con-

cerned with slavery in comparative terms. The matter of what has been created remains unresolved. THE CULTURAL PROBLEM

In order to come to terms with these issues we must turn to the cultural problems posed by the release from slavery. It is by cultural and symbolic means that the question, What has been created? is answered and at the same time the nature of the transaction is resolved. Since the slave is natally alienated and culturally dead, the release from slavery has certain implications in terms of symbolic logic. As enslavement is life-taking, it follows logically and symbolically that the release from slavery is life-giving and life-creating. The master gives, and in giving he creates. It must always be

the case that the master gives up something, so that the slave may gain something else. The master would seem to gain nothing. Hence he incurs a loss. What results from this deliberate loss is a double negation: the negation of the negation of social life, resulting in a new creation—the new man, the freed man. Manumission, then, is not simply an act of creation: it 1s, rather, an act of creation brought about by an act of double negation initiated by the freely given decision on the part of the master to part with something— his power—for nothing. To be sure, the slave often gives something too: his redemption fee. But not only does this not pay for what the master loses— his power—it is not even possible for it to pay for anything, for whatever the slave gives already belongs to the master. Hence, even when the slave pays, he is really not paying for his freedom. It is usually conceived of as making a gift offering in gratitude for the master’s freely given decision to release him from slavery, however that release is arranged. THE THEORY OF GIFT EXCHANGE

The whole complex of ideas and interactions involved in the release from slavery amounts to a classic instance of the anthropology of gift exchange. A

distinguished line of anthropologists from Marcel Mauss and Bronislaw

212 Slavery as an Institutional Process

Malinowski through Raymond Firth and Marshall Sahlins has demonstrated the enormous significance of gift exchange not only as a utilitarian premarket means of exchanging and redistributing goods, services, and other resources, but as a means of striking new social compacts and reconfirming old ones.’ Gift exchanges range in complexity from simple diadic relationships to intricate systems of interaction involving many persons over long periods. Mauss uses the term “prestation” to define such systems, and the most elaborate ones, having implications for the entire social order, he calls systems of “total prestation.”® Mauss’s arguments may be formalized as follows. There is a utilitarian

component, which refers to the material exchange of goods and other resources resulting in a net balance or imbalance as the case may be. And there is an ideological component, which is the conscious rationalization and moral expression of what is actually going on. The ideology of the prestation departs in varying degrees from its reality. In some cases the variance may be minor and quite transparent. It is often held, for example, that a gift is freely given, without any thought of reciprocation, when in fact everyone knows that all kinds of obligations are established by the act. It may even be held that the value of the gift offered is of no importance whatever, only the spirit in which it is given. In this regard modern societies are considerably more hypocritical and mystifying than primitive ones, since in the latter the exact value of the gift was frequently spelled out in great detail. At the other extreme, the ideology of the prestation may interpret the actual transaction

as the very opposite of what is actually happening. The recipient may be looked upon as the person who benefits most, and the giver as the person who has magnanimously given far more than he can ever hope to receive; in reality, the very opposite may be the case. Manumission, we shall soon see, is one such kind of prestation. The symbolic component establishes the gift exchange as “a social compact.” It synthesizes the ideological and the utilitarian components as counterpoised elements in a single ritual process. Ritual not only mediates between the two other components, but further mediates between the specific interaction and the total system of interactions that make up the entire exchange system. In this way each subprocess of prestation is given social and moral significance. Speaking of the category of prestation in which there is what he calls “balanced reciprocity,” Sahlins has written that “the striking of equivalence, or at least some approach to balance, is a demonstrable forgoing of self-interest on each side, some renunciation of hostile intent or of indifference in favor of mutuality. Against the preexisting context of separateness, the material balance signifies a new state of affairs ... Whatever the utilitarian value, and there need be none, there is always a ‘moral’ purpose.””’ Mauss’s major contribution was his analysis of the dialectic of the ritual process that transmits and generalizes the utilitarian and ideological compo-

Manumission: Its Meaning and Modes — 213

nents of the gift exchange. All prestations, he showed, involve three kinds of obligations: there is an obligation to give presents, which in many precapitalist societies was not only morally motivated but may have been a material necessity. This, however, determines that there is a counteractive obligation

to receive, since not to do so might not only be an insult to the giver, but (again more in premarket systems) would be an unpardonable break in the chain of exchanges that establishes a moral order, substitutes for a market, and also ensures redistribution. But to fulfill one’s obligation to receive can only be resolved in the synthetic obligation to repay the debts imposed by the receipt of gifts, for “the gift not yet repaid debases the man who accepted it.”* If the relationship is an ongoing one, it is clear that repayments both complete and initiate a cycle of gift exchanges in a continuous dialectical progression that moves forward lineally for the two persons interacting, but concurrently spreads out laterally to all persons interacting in the total system of prestation—in other words, to the community at large. Later scholars have considerably refined Mauss’s analysis, although in some instances the profundity of Mauss’s original insights has been missed,

partly because of the highly elliptical style in which he wrote. Firth and ‘Sahlins,’ in particular, have rid the original analysis of its mystical aspects, though in strikingly different ways. Firth emphasizes that Mauss’s analysis is, in its entirety, more applicable to premodern societies than to modern ones; in the latter, notions of political morality, not to mention the operations of the market, either obviate or impose deliberate restraints on the obligations to give, receive, and repay. Firth’s distinction between different kinds of gifts is also useful, especially his categories of the earnest, a gift that is “an indication of what is further to come, or what may further come if certain conditions are met by the recipient,” and the token, which may serve symbolically either as “an index of commitment” or as a rejection of any such commitment.’° Both Firth and Sahlins emphasize the fact that gift exchanges are frequently asymmetrical. Sahlins’ distinction between “balanced” and other kinds of reciprocity may be useful for schematic purposes. It seems to me, however, that Mauss was only too aware of these distinctions. THUS WHILE IT IS TRUE that on the utilitarian level prestations vary consider-

ably in their degree of symmetry, on the ideological level all exchanges are interpreted as balanced and fair. This is true even in an advanced capitalist society, where the ideology of the free market in labor insists on a “balance” and ‘‘fairness”—which in reality is anything but balanced. I have already discussed in an earlier chapter Marx’s analysis of the way in which the fetishism of commodities acts as a powerful ideology of equal exchange in the face of the actual inequality of exchange involved. I suggest that Mauss’s analysis of gift exchange is the precapitalist counterpart of such an ideology.

214 Slavery as an Institutional Process Once he established the utilitarian dialectic of giving, receiving, and repaying, Mauss shifted the focus of his analysis to the wider dialectic of generalized exchange in premodern societies. One aspect of Mauss’s analysis that has received more attention from

students of comparative religion than from mainstream anthropologists concerns gift exchange between man and god. This takes the form of sacrifice. As Mauss notes: ‘Sacrificial destruction implies giving something that has to be repaid.”'' Mauss argues that such sacrificial exchanges between man and gods (or spirits of the dead) may have been the very first prestations. The gods and spirits of the dead “in fact are the real owners of the world’s wealth. With them it was particularly necessary to exchange and particularly dangerous not to; but, on the other hand, with them exchange was easiest and safest.”’'* Sacrifice is not just the earliest but the fullest realization of prestation systems, “for the gods who give and repay are there to give something great in exchange for something small.” | have added emphasis to the last passage, for it clearly indicates Mauss’s understanding of the asymmetry that underlies the ideological and symbolic symmetry of all gift exchanges. Clearly, men do not outwit the gods in getting something great for something small—-they only think they do. In the end, however, everything returns to the gods. The asymmetry works on their behalf. What Mauss is suggesting here—though he is at his most elliptical—is the fascinating fact that both parties may think that they have benefited disproportionately, and indeed may have done so, given their respective points of view. What one gains is wholly relative to one’s status, aspirations, and needs. And who is to say that the person who thinks he has gained has not in fact gained? It is in this sense that all prestations are reciprocally balanced.

Finally, Mauss makes the important observation that there is often a close relationship between gift exchanges among men, on the one hand, and among men and gods on the other, the two reinforcing each other. He notes of the potlatch of the northwestern American Indians: “It is not simply to

show power and wealth and unselfishness that a man puts his slaves to death, burns his precious oil, throws coppers into the sea, and sets his house on fire. In doing this he is also sacrificing to the gods and the spirits who ap-

pear incarnate in the men who are at once their name-sakes and ritual allies.”'’ Mauss sees the origin of almsgiving in prestation, which serves the double purpose of being an offering to the gods while perpetuating prestation among men. | shall have more to say on this tantalizingly brief “note” of Mauss’s shortly, for, as we shall see, the origins of manumission itself can be explained in terms of these double-purpose man-god/man-man prestations.

The Rituals of Redemption To return now to release from slavery, | want to examine a sample of release ceremonies among a wide range of peoples in order to show how the prin-

Manumission: Its Meaning and Modes — 215

ciple of prestation permeates them all. After this preliminary review, | shall examine the data in a more detailed and systematic way, to isolate the main patterns of release within and between the different slaveholding societies. The Kongo community of Mbanza Manteke (in Zaire) during the nineteenth century is a good case with which to begin, for it is typical of the extreme category of societies in which the inalienability problem was insoluble. Once a person became a slave, there was no way ever of releasing him from this condition, even though he might have been released from a particular master or even from his master’s community. “A slave is a slave forever” was the rule. The release from a particular master and his clan “was always possible providing that the slave’s owners and his original clan would cooperate; the standard ritual required that the slave be marked with chalk as a sign of redemption and that a pig called ‘the pig of rubbing with chalk’ be transferred to the ex-owners. But the ‘redeemed’ slave was then considered to be the slave of his original clan and could never recover first-class citizenship—that is, access to authority.”' The release ceremony was obviously a gift exchange. It is significant, however, that the exchange was not between master and slave but between the master and the slave’s former clan members. The transaction involved another transition for the slave—hence the chalk mark, a common symbol of such transitional conditions in Africa. The chalk is also a symbol of death, which the slave continued to experience socially. A significant number of other primitive societies hold the view that social death, like its physical counterpart, is irrevocable, even for the slave returning to his natal clan.'° At the other extreme, and typical of all advanced slaveholding groups, is the Roman law of postliminium, by which the slave who returned home was fully restored to his old status. Most societies fell within these two extremes in providing some possibility of release from slavery. Typical of kin-based societies were the Kerebe of Tanzania. In addition to the usual compensation (one cow), slaves had to send a gift of a hoe or a goat to the omukama, or chief, and “the person being released from servility was taken to a crossroad where his head was shaven to symbolize the loss of servility.”'® We have already seen that the shaving of the head is a common symbol of transition and death and hence of enslavement. Here the shave again symbolizes transition. It also symbolizes death, only now the double negation implied in the death of the social death of the slave. This is reinforced by the crossroads, another symbol of transition and a common location for ritual events the world over. Nor should we neglect the more obvious symbolic meaning of the crossroads as a sign of free choice. The Kerebe and the Romans of the principate stand at two extremes of sociocultural complexity in the premodern world. Yet it is significant that among the Romans freedmen played a disproportionate role in the Lares cult, which was closely associated with crossroads and intersections. Indeed, when the genius of Augustus became the object of religious honors and the Lares cult adapted for this purpose

216 Slavery as an Institutional Process (becoming the Lares Augusti), it was freedmen assisted by slaves who were in charge of the rites. J. H. W. Liebeschuetz’ explanation for the prominent role of freedmen in this cult in both its older and its Augustan forms is unpersuasive—that it “was appropriate since freedmen formed a high proportion of the population of Rome.” Clearly their role must be accounted for in terms of the deep and widespread symbolic association of the emancipated with the cult of the crossroads, And it becomes evident that the shrewd emperor had adapted this particular cult in order to make a powerful symbolic statement to his people—that loyalty to his person and his genius was identical with freedom."’ Inasmuch as the human head looms large in the ritual of enslavement, we are not surprised to find it in prominence in the ritual of redemption. One of the most archaic of the Frostathing laws of Norway mandated that “if a slave takes a tenancy or farms for himself, he must make his freedom

feast, each man with ale brewed from three measures, and slaughter a wether—a freeborn man is to cut its head off—and his master is to take his neck-redemption off his neck.” Significantly, the Norse word for free, frials, is derived from frihals, which means “free-neck.”'* In ancient India during the period of the Tipitaka the manumission ritual was simple but potent:

“A master desirous of freeing a slave would wash his head and declare him to be a bhujissa, a freeman.” Sometimes, however, the master simply ordered the slaves to “wash their heads themselves and consider themselves free.” A further development took place in India between the second century B.c. and the fourth century A.D. Slavery had been considerably mitigated, and manumission and its implications were both significant and well-defined. The ceremony, accordingly, was more elaborate. “A master desirous of manumitting his slaves will take away a jar full of water from the shoulder of that slave and will break it. He will then shower some parched grain and flowers on the slave’s head and repeat thrice, “You are no longer a dasa.’ This act symbolized the cessation of his duty of carrying water. With this cessation all servile duties were discontinued for him.”"” I suggest that this is only the most manifest layer of meaning in a complex, evocative rite. Water is the symbol of purification and of regeneration in India, as in many other cultures. The breaking of a jar of water clearly implies both the destruction and the rebirth inherent in the act of manumission. There is the same binary opposition in the act of showering the slave’s head with parched grain and flowers. Rice, of course, is used in rite-of-pasSage ceremonies all over Asia, especially as a fertility or birth symbol. But note that here the rice is parched, another striking opposition, for parched rice is dead. Thus at one stroke have been expressed both the idea of death and the idea of renewal. The social death of the slave is destroyed as in a broken jar or a handful of parched rice. In both there is the simulation of a

Manumission: Its Meaning and Modes 217 sacrificial act: the broken body of the jar, the burned rice. The slave dies as slave. Death is negated and there is the hint of renewal, of new bloom, in the flowers thrown on the slave’s head. It is among the medieval Germanic peoples, especially the Scandinavians, that we find the idea of manumission as gift exchange most richly expressed in symbol and ideology. Agnes Wergeland writes of the German master’s view that “freedom was, after all, more in the nature of a gift than a purchase, and was also most of the time called by that name.””° To be sure, he did receive a payment from the slave, but “to the master, the fee would not compensate for the loss of a permanent laborer, and the slave could thus not be said to have really paid for his freedom.” In modern capitalistic slave systems, even though slaves paid dearly in one way or another for their freedom, that freedom itself was still regarded as a gift from the master or mistress. The cartas, or letters of manumission, written by masters during the colonial period of Brazil are very revealing. Stuart B. Schwartz found that “statements within the letters—and the pride with which masters granted manumission even to the old and infirm—indicate that slaveowners saw the act of manumission as a charitable gesture no matter what its conditions or terms.””! In the U. S. South, where the manumission rate was one of the lowest of all slave systems, those few masters (mainly in the cities and border states) who allowed their slaves to buy their freedom “understood that freedom was the greatest gift they could give their slaves, and they consciously used it as a mechanism of control and a means of encouraging divisions among blacks.” The premodern world exhibits some striking parallels in the rituals of release. The “English mode,” for example, consisted in “transferring the slave from the hand of the master to that of another freeman, who manumitted, as a symbol of the separation from the lord. This was always done in the presence of the assembled free. The liberated was then shown the open road and door to signify that nobody could restrain him, and he was given a freeman’s sword and spear.”” More common was the ceremony found among the Langobards: “The slave amid clashing of arms in the assembly [gairpinx] was passed from the hand of the lord to other freemen till the fourth was reached. He then declared him free and completed the act by leading him to a crossway, bidding him to be at liberty and to go where he wished. The slave was then given arms and was henceforth a full free Langobard.”*? The Germans, like many tribal peoples, required the presence and confirmation of the entire local community in all acts of partial manumission; where full manumission was involved, the king’s presence was required among certain groups. Among the Franks, “the slave was liberated before the assembled freemen by the king [originally the lord] knocking a penny from the slave’s hand so that it flew over his head. This was a sign that his services and dues were dispensed with.” The Germans distinguished between different degrees

218 Slavery as an Institutional Process and stages of manumission. Perhaps the most characteristically Germanic ritual of redemption was that known as the liberation beer. This came when a partially free slave wished to complete his freedom. Wergeland writes that “this was a festivity prepared by the [manumitted] slave to celebrate and at the same time to make public his release. Here, in the presence of company sufficiently numerous to witness the act, he was to offer the lawfully demanded fee, which in this case must be nominal since it represented only one-fourth of what was looked upon as his suitable market value.”** Closely related to this rite but more elaborate was the ceremony performed for the child of a master and a slave at the point when he was to become a full and free member of the community. At a social gathering an ox was slaughtered and a large quantity of beer brewed. A shoe was then made from the skin of the right forefoot of the ox, and the formal legitimation of the child began: first the father, then the child to be completely manumitted, then the nearest heir, followed by the rest of the family, all stepped into the shoe, “each pronouncing at the same time the appropriate formula which indicated the particular meaning of the ceremony.””” A final example is the manumission ritual of early Babylonia, which involved the cleansing of the forehead of the slave and then turning him or her to face the rising sun. The Ugaritic custom was to pour oil on the slave’s head. Manumission was often conducted in the presence of a priest or judge, and the fully manumitted slave was made “like a son of the city.” The exact nature and meaning of the ceremony of cleansing the brow is a source of some controversy. G. R. Driver and J. C. Miles claim that the cleansing of the brow (in which a slave-mark was literally removed) and the turning of the slave toward the rising sun were “two parts of a simple religious rite, [which] usually took place in the temple of the sun god Samas.” It is generally agreed that the cleansing ceremony involved the use of water and was a form of purification, water being very important in the ritual of Babylonia. However, scholars are not in agreement on whether the ceremony also involved dedication to a god, as in ancient Greece.*® All these ceremonies emphasize the same themes—the communal nature of the act of manumission, the fact that the manumitted gains power that he formerly lacked (represented for example in the handing over of weapons), the acquisition of the capacity to compete for honor (represented in the German case in the passing of the individual from one freeman to another or in the ritual of stepping in the same shoe), the attainment of will and autonomy (represented by the crossroad), the negation of social death and transition to a new status, and most important, the concept that the master or his clan makes a gift of freedom and that what the slave pays is merely an offering, a gift exchange. No slave, of course, took all this literally. Still, the ceremony was for all of them deeply meaningful and the idea of giving, receiving, and repaying was gratefully accepted. The ideology of gift exchange

Manumission: Its Meaning and Modes 219 and the symbolism of free choice were one set of things; the reality of what the slave actually paid, and was required to continue paying, was quite another. To understand the totality of what transpired in the various rituals of release, we must now examine the data more systematically.

The Modes of Release There were various modes of release from slavery, and most societies at any given time employed several of them. Manumission was not itself a constant: in a given society what a slave achieved through manumission varied. Some slaves achieved full manumission at once, others attained it over time,

still others remained for the rest of their lives in a twilight state of semimanumission. The different modes of release reflected such differences in the kinds of manumission. The modes of release varied too according to which party initiated the release and according to the factors that motivated both slave and master. In some cases, both wished to remain closely tied to each other; in other cases,

the slave wished to be as far removed from the master as possible, to the point of returning home or migrating to another more favorable society; and in others, it was the master who wished the ex-slave to remove himself. Different modes of release might be used depending on the expected outcome.

- The final variable was the legal and cultural idiom employed in rationalizing and legitimizing the release. Special problems perhaps existed, which had to be resolved before the master was capable of releasing his or

her slave, even if he were willing. For example, minors were not usually permitted to manumit in ancient Rome and the French West Indian colonies during the eighteenth century, and in many societies mistresses were forbidden to manumit their male slaves in order to marry them. Exceptions were usually found in special modes of manumission that circumvented the legal prohibition. Or the mode could simply be a cultural survival, a functionally obsolete method that persisted from a previous period when it was meaningful. This was true of manumissio censu (a form of political manumission) during the late republic and the early imperial period of ancient Rome. With a few isolated and culturally peculiar exceptions, the modes of release throughout the slaveholding world were basically of seven types: (1) postmortem; (2) cohabitation; (3) adoption; (4) political; (5) collusive litigation; (6) sacral; and (7) purely contractual. POSTMORTEM MANUMISSION

The postmortem mode was one of the most widespread, and most likely was one of the earliest modes. By it I mean the release of the slave at the death of

220 Slavery as an Institutional Process his master, whether by written will, or by verbally expressed desire, or by heirs on behalf of the deceased master. There are many reasons why postmortem manumission was so common and so early to develop. The master who freed his slave after his death incurred no personal cost to himself. There was, of course, a cost to his heirs, but this was more than compensated for by the second factor accounting for the popularity of this mode: it was one of the most effective means of motivating the slave to accept his role and effectively perform his assigned task, whatever it was. The mere possibility of postmortem manumission motivated all slaves in a large household, even if eventually only one or two were manumitted.”’ The origins of this mode and the fact that 1t was one of the earliest forms of manumission among most peoples was closely related to the primitive practice of using slaves as sacrificial victims and as gift exchange among persons. We have seen that slaves were killed for three main reasons: as sacrifice to the gods or ancestral spirits; to accompany the master in the afterlife; and to display the master’s prestige, power, and wealth, either during his life or at his death. The three uses are clearly interrelated: the same gift can serve the double purpose of an exchange among men and between men and gods. In the same way that the slaughter of slaves at potlatches was as much for the benefit of impressing men as for a sacrifice to the gods, so the burial of slaves with their masters may have served as much for accompaniment of the master to the world of the dead as it did for a sacrifice to the ancestral spirits and gods whom the master was about to join. Indeed, such mortuary sacrifices may even have served a third function: to impress the living with the honor, prestige, and wealth of the deceased. Illustrative of such multivocal gift exchanges between men and gods, combined with spiritual accompaniment in the afterlife and a final display of honor and wealth, is the beautiful and brutal story of the Icelandic princess Bryndhild, who did not wish to live after her beloved Sigurd had died. She had her eight male and five female slaves killed, and before committing suicide herself, she ordered: “Bedeck the pyre with shields and hangings, variegated Welsh [foreign] cloth and Welsh corpses. Let the hun [that is, Sigurd] be burned on the side of me; on the other side, my servants with their precious ornaments and two hawks.” The Welsh corpses were “to honor the dead.” She added: “Then will our procession not appear mean and poor, for it shall be followed by five female thralls and eight male thralls of gentle birth reared by me.”** Among a great many peoples who sacrificed slaves coexisted the practice of freeing some slaves on the ceremonial occasions when others were sacrificed. What this immediately suggests is that the killing and the freeing of a slave were symbolically identical acts—an identification that makes a good deal of sense. To own a slave, after all, meant simply that one had exclusive proprietary powers in him; and such powers were equally destroyed whether

Manumission: Its Meaning and Modes 221 the slave was killed or given away or freed. The principle of gift exchange was served by all three forms of proprietary destruction: in the first, the killing of the slave, the exchange was with a deity or spirit; in the second it was with the slave himself; in the last it was with a third party. Furthermore, giving the slave to himself sometimes also served the dual exchange function

between master and slave and between master and god. | The comparative data allow us to observe this process of substitution during the transition stage when sacrifice and emancipation both were practiced, as well as later, when human sacrifice of the slave had been totally re-

placed by manumission. Let us examine the process among three widely separated groups: the Garos of India, the Toradjas of the central Celebes, and the Indians of the northwest coast of America. Several observers have established that human sacrifice of slaves was common “in olden times” among the Garos. When this practice was abolished, it was replaced by the sacrifice of an animal—a bull in some areas, a goat in others—at the cremation of the corpse. However, in direct reference to the earlier practice of sacrificing a slave, “a living slave [was] tied to the leg of the corpse from the day of the decease to the hour of the cremation. He or she was then released from further servitude.””” Among the Toradjas the development of manumission from the sacrificial murder of the slave at the death of the owner is readily observed. Of its earliest form we learn that at the death of a prominent person a slave was bought from a neighboring tribe for the sacrifice, or if the tribe of the deceased was at war, a head-hunting expedition would go in search of heads. In a second development that began later but ran concurrently with the first, a slave was designated the tandojai, the person who took food to the deceased and guarded his body from the werewolf to prevent the latter from eating it. The tandojai would speak to no one during this period, and he was free to

take food for the dead master from anyone. It was said that “the tandojai lives like a death soul who now has a death soul in service.” Sometimes instead of assigning a slave as tandojai, a piece of his ear was removed and the blood running from it was rubbed on the coffin. Where the slave was made a tandojai, however, he was always freed after the funeral services; significantly, he led a solitary life and people feared him. In other areas, especially after the arrival of the Dutch, a later development seems to have been the simple guarding of the grave by a slave who was then freed a hundred days after the funeral. The Toradja case clearly indicates that such substitutions

were not due to European influences and prohibitions, since they existed alongside the continuing practice of human sacrifice and before Dutch contact.” We find the same striking evidence of the identification of sacrifice with manumission among many of the American northwest coastal tribes where manumission at ceremonial occasions, especially mortuary ones, was often

222 Slavery as an Institutional Process the only mode of release from slavery. Among the Tlingit, slaves were both sacrificed and freed at potlatches and at feasts for the social reception of children; for example, the slave girl who attended her young mistress during the period of her seclusion before she was brought out into society was freed after the burning of the young mistress’ old clothes.’' A slave who disappeared at the cremation of his master was usually considered free, as was the slave who dressed the master in his funeral robe.** Among the Nootka, it was the custom to give away some slaves, in addition to killing some, upon the death of a chief.*”

It does not seem unreasonable to speculate that in many instances of postmortem manumission the identification of release with the sacrificial murder of slaves has survived. We know, for example, that among the Aztecs postmortem manumission was the most common mode of release, and given the important role of the ritual sacrifice of slaves in this society, it is certainly not unreasonable to assume that postmortem release was a con-

tinuation of the earlier practice of identifying the killing and freeing of slaves as parallel symbolic acts. The later Christian practice of testamentary manumission substituted easily for this tradition.** In much the same way, the Islamic practices of testamentary manumission seem to have replaced— without an intervening period of pagan postmortem manumission—the earlier Wolof and Sereer practice of human sacrifice at the death of the master, which continued as late as the sixteenth century.” What of the origins of testamentary manumission, the postliterate form of the postmortem mode, in Christianity and Islam? The Western mode of testamentary manumission preceded Christianity by many centuries and was simply adopted and sanctified by the early church. The problem, then.

is to explain the origins of the practice in pre-Christian times. Unfortunately, there are few data, so it is impossible to do more than speculate. Postmortem manumission may well have developed as a substitute for human sacrifice, in much the same way that it emerged among many contemporary primitives. We have already seen that the custom of human sacri-

fice at the death of important persons was practiced among the primitive Europeans, as it was in the ancient Near East, the Orient—and, indeed, all other parts of the world. Primitive Rome was no exception. The circumstantial evidence is very suggestive. Manumission in testamento was one of the earliest modes of manumission in ancient Rome and must have long preceded the Twelve Tables, since the Law of the Twelve Tables, which refers to it, takes the form of a confirmation of existing practice. In early times the will was ratified by the primitive assembly. David Daube notes that “the element of public control would be present.” He argues forcefully that manumission “testamento and vindicta must have preceded censu.’’ The same arguments also point to the historical primacy of manumission testamento over vindicta. As Daube points out, “In the history of law private actions authorized by the community preceded ‘community-

Manumission: Its Meaning and Modes — 223 initiated or state actions.’” Furthermore, “a head of a family, a pater familias,

was stronger and more independent in the primitive, small, loosely organized society than later in a developed and numerous state.”*° Of all the modes of manumission, testamento was the most private. Partly for this reason it did not require any legal convention, only social confirmation. Thus, the very legal maturity of manumission vindicta—maturity in the sense of employing a rather sophisticated legal fiction, to be discussed below—suggests that it was the less archaic of the two. This was also the basis of Appleton’s inference that manumission in testamento was the oldest mode in ancient Rome: an inference that Buckland rejects, claiming that “the contrary

conclusion seems more reasonable.”*’ Buckland gives no reason for this contention; perhaps he assumed that the existence of a legal fiction is indicative of a more archaic form. If so, he was wrong: fictitious litigation is the surest sign of legal maturity, as anyone acquainted with primitive law will attest. The comparative, circumstantial, and internal legal evidence all point to

the historical primacy of manumission in testamento in primitive Roman law. Whether or not this earliest form of manumission developed, in turn, from an earlier custom of symbolically identifying the sacrifice of slaves with their release must remain a matter for speculation. Consider, however, the evidence we do have. We know that the primitive Romans kept slaves, and that the institution of slavery was of great antiquity. We know that the

primitive Europeans performed human sacrifice at the death of slave owners, and we have no reason to believe that the Romans were any different on this score. Indeed, apparently under Gaulish influence, there was a revival of the practice of human sacrifice in 225 B.c. when several captives, including two Gauls, were buried alive. Instances of human sacrifice occurred for the next three hundred years, although outlawed by the Senate in 97 B.c. The primitive practice of substituting an animal for a human being when human sacrifice was no longer allowed is attested in historical times, most notably the offering of a goat as surrogate for a human victim in the cult of Vediovis.*®

Whatever its origins, testamentary manumission was the most popular mode of release from slavery throughout Roman antiquity. R. H. Barrow summarizes the main reasons as follows: “Manumission by will had advantages: it retained the services of slaves to the very last moment in which their owner could use them; it kept the slaves in a suspense of good conduct to the end. Trimalchio made no secret of the provisions of his will: ‘my object in making them known is simply that my household may love me as if I were

dead.’ The manumitter departed from life in a comfortable glow of selfrighteousness, which he may have earned by this one deed. He could rely on the grudging gratitude of those who may have hated him, and could trust that a suitable gathering of mourners would lend more than mere respecta-

bility to his funeral.”

224 ~~ Slavery as an Institutional Process

There was, however, a great deal more involved than Barrow imagined. Such manumissions may have been a vestigial gift exchange with the gods, a last sacrificial repayment before departing this world (though in a highly symbolic way). It is important to emphasize that there was never, either in primitive or classical times, any notion of piety involved; this would have

been alien to Roman religious conceptions, as it would have been to all other pagans. Second, such manumissions were forms of gift exchange be-

tween the dead man and his successors, for anything that enhanced his prestige must have enhanced the prestige of his successors. Significantly, sacrifice of slaves used to perform very much the same function. T. C. Ryan,

in his study of the economics of human sacrifice in Africa, found that the most likely model to explain the incidence of sacrifice was that of gift exchange, even where enhancement of the donor’s prestige was also involved. He concludes: “Sacrifices at the funeral obsequies of an important person, while sending his favorite slaves to attend him in the hereafter, also asserted the wealth and power of his successor.”*” Exactly the same could be said for the release of slaves at the death of the master. There was a third gift exchange involved, again largely on the symbolic level, though powerfully so: that between deceased master and slave. The master’s death was the occasion for the release of the slave. It is a short step from this to the position that the master had died so that the slave might be

free, that is, might be born again into social life. This placed the ex-slave under the deepest possible obligation to repay the gift of the master by honoring him for the rest of his life, and of course also honoring and serving his

successors. The slave’s manumission may also have had an even deeper symbolic meaning. The slave was an extension of the master’s self, a view given legal expression in many slave codes (one of the most noteworthy being that of early modern Russia).*' The freeing of the slave at the death of the master may well have had a death-defying and recreative meaning: the master’s spirit resurrected in the living person of his favorite surrogate. If this was so, it placed the manumitted slave under an even greater obligation: he had to be not only grateful, but a faithful vehicle of the deceased man’s spirit. Whether such symbolic meanings existed among the late republican and early imperial Romans is problematic. The symbolism of Christianity, and its enormous success in the Roman world, strongly suggest that such meanings were present in the later empire. Certainly many primitives held such views: among several tribes, for instance, the slaves freed at the death of the master were greatly feared.

There was one last critical development in the symbolic meaning of postmortem manumission: the radically new conception of the relation between man, god, and the cosmos. Robert N. Bellah and others have shown how the development of the great world religions entailed a collapse of the cosmological monism of both primitive and archaic religions. The world was

| Manumission: Its Meaning and Modes — 225 no longer seen as a single cosmos in which men and gods participated in sacred and profane fields, but rather as two sharply polarized cosmos—one

centered on the present world, the other on the life hereafter. The major ethical implications of this cosmological dualism were, on the one hand, rejection of the present world as evil and man as inherently unworthy, and on the other hand, emergence of the idea of salvation as the central religious preoccupation. Ritual and sacrifice remained prominent but, as Bellah tells us, they took on a new significance; they were no longer directed so much at fulfilling obligations to the gods by establishing harmony with the cosmos, but at the primary goal of salvation, that of saving man from his original evil and sinfulness and ensuring a place in heaven rather than hell in the other cosmos, which is entered at death.” How did this development influence the interpretation of manumission? The impact was tremendous, but the various world religions came to a recognition of the relationship in different ways. Releasing one’s slave from slavery eventually was seen as a pious act that would be rewarded in heaven: it was almost an ethical imperative in any transcendental ethical creed that emphasized humanism, individualism, and salvation, as did all the major world religions. And indeed, they all came to this view in time. To begin with Christianity, what is immediately striking is the length of time it took to realize this imperative, and the circumstances under which the realization came about. Until the end of the Roman Empire, Christianity had no influence whatever on the meaning and motivation of this form of manumission. Even after it had become an official creed, Christianity remained indifferent to manumission in general. To be sure, the church, from as early as the third century, encouraged the ransoming of captives, but this was motivated by a horror of Christian souls being enslaved by heathens, not by any aversion to slavery per se. Manumission in the church was encouraged from the fifth century on, but the emphasis was on manumission inter vivos. The objective was to invest most manumissions with Christian

ceremony. As late as the sixth century the church had yet to develop any notion of the virtuousness of manumission, and there was no conception whatever of the special virtue of testamentary manumission. It was not until the start of the seventh century that we find the first forcefully articulated theological statement that manumission in general was an act of piety; it came from Saint Gregory the Great, who took to their logical conclusion the reservations sounded earlier by Chrysostom and Cyril of Alexandria. Even so, there was no attempt to single out testamentary manumission as the form that was most expressive of piety.*’ All of this changed dramatically during the ninth and tenth centuries. At last manumission was not only encouraged but viewed as good for the soul. Piety and salvation were especially ensured by testamentary manumission. The following for-

mulaic statements attached to testamentary manumission are typical:

226 Slavery as an Institutional Process “While Almighty God gives us health in this life, we ought frequently to think of the salvation of our souls—and so I, for the good of my soul, and to break off my sins that God may pardon me in future, have released my slave —__— and have given him his peculium.” Or the following: “It behooves every man in this life to think of the good of his soul, and so I, in God’s name, having regard to God and the redemption of my soul, manumit etc., etc.” Or finally: “He who releases the bondservice due to him may hope that he may receive a reward in the future from God, and so for my eternal

retribution I manumit, etc. etc.” How do we explain this extraordinary change in the religious perception of manumission? Our concern here is not with the influence of Christianity on the frequency of manumission, but on its rationale or ideology. Why did Christianity wait seven hundred years before reinterpreting the meaning of manumission, and nearly nine hundred years before encouraging testamentary manumission as a form of piety that was good for the reception of the soul in the afterlife? The answer is simple. Early medieval Christianity was fully involved with the problem of converting the heathen peoples of both central and northern Europe, especially the Germanic tribes. The sacrifice of slaves, as well as their manumission at the death of the master, were wellestablished pagan practices among these primitive Europeans. And the advocates of the new religion were determined to stamp out all forms of paganism, including the sacrifice of slaves at their master’s death. Hence the manumission of slaves at their master’s death, which already existed as a sacrifice substitute, was reinforced and encouraged; but where the pagan practice of postmortem manumission was meant to honor the departed in

this life or serve as a final offering to the pagan gods, the meaning was changed to securing the salvation of the soul of the departed in the next world. The shift of meaning was easily understandable to the pagans, since it was already partly for this purpose that the slaves were sacrificed. By assimilating the pagan meaning of postmortem sacrifice to the Christian mean-

ing of piety and redemption of soul through the encouragement of testamentary manumission, the church was able to achieve two critical objectives simultaneously: it kept the pagan practice of postmortem manumission, but

changed its meaning to the Christian one of salvation of the soul in the afterlife; and it kept a good part of the meaning of the pagan practice of postmortem sacrifice of slaves—assistance of the master in his passage to the hereafter—-while abolishing the practice itself. Although this transformation is easily documented from the history of all the central and north European peoples, it is best observed in medieval Sweden where Christianization was “gentle and tentative” and where the abundance of testamentary evidence shows clearly how the church introduced “an important second incentive for emancipation: to free a slave was meritorious in the eyes of God and contributed toward the earning of salvation.”*”

Manumission: Its Meaning and Modes = 227 Much the same process occurred in the relation of Christianity to modern pagan peoples. The same reinterpretation, for example, quickly took place among the Aztecs, and also among the tribes whose practices of postmortem manumission and sacrifice have already been discussed. But acculturation is always a two-way process. Christianity may have gained more than it lost in this neat piece of theological reinterpretation; still, it would be wrong to assume that the priests had it all their own way. There was compromise. The primitive mind did bend the faith, if only in a small way, to its own manner of thinking. We may speculate that, to the converted, primitive testamentary manumission became a powerful binary symbol. The socially dead slave was made into a socially alive person. The master’s death became the occasion for a life-creating act, which pleased not only the men he left behind but the new god he was to meet. The element of sacrifice to the god persisted, only in its more sophisticated form of sacrifice in property. The

element of gift exchange was still present—a life here on earth was exchanged for a life there in heaven. Christianity had already anticipated this interpretation: the Crucifixion was an enormously powerful and multivocal symbol. It may have been altogether too powerful and too primitive for the sophisticated urban world of civilized Rome. Thus Paul, when asked to give his views on manumission, made the following paradoxical reply: “Regarding the matter of which you wrote me, the slave who is called by the Lord is a freedman of the Lord, likewise the free man who 1s called is a slave of Christ. You have been bought for a price, be not slaves of men. Let each man, brothers, remain beside God in that status in which he was called.’’*° This theological obscurantism left the early fathers utterly confused. Paul himself may have been none too clear about what he meant. Much the same process took place in the spread of Islam and its relations, first to the pagan Arabs and later to the other pagan peoples it converted to its faith. The Arabs of Muhammad’s day had, like the Romans of

the early empire, already abandoned the human sacrifice of slaves and practiced only postmortem manumission. But like the Romans and Greeks, such manumission was directed mainly at the enhancement of the deceased master’s good name and honor. We know that Muhammad and his followers accepted the existence of slavery as part of the social order, but much more quickly than Christianity, Islam sought to humanize the institution. To this end, Muhammad not only encouraged manumission inter vivos, but established as a cardinal principle the idea that manumission, especially the testamentary mode, was a pious act that was good for the master’s soul. Indeed, Islam went even further. Not only was the master enjoined to free his slaves in the hope of reward in the afterlife, but his heirs were enjoined to do so, for the “freeing of a slave on behalf of a person who is dead is profitable for the dead.”’*’ And there is abundant evidence that Islam, like Christianity among the Germanic pagans, reinterpreted the traditional practice of the

228 Slavery as an Institutional Process sacrifice and release of slaves among the pagan peoples of Africa and Asia it

proselytized. We have already cited the case of the Wolof of the Gambia, who sacrificed slaves as late as the sixteenth century, then replaced this practice with the Islamic doctrine of manumission as a form of piety and redemption for the maltreatment of slaves. THE COHABITATIONAL MODE

This mode of manumission, by means of marriage or concubinage, was the most common form among lineage- or kin-based preliterate peoples, and it was also the most common mode in the Islamic world, especially in Africa and the Middle East. The reasons are not hard to find. It was easy for men in most precapitalist societies to identify the status of a female slave with that of free concubine or junior wife. The difference in status usually had no material consequences for the woman, although it did for her children. Manumission of, and cohabitation with, female slaves was not only allowed but actively encouraged in Islam. Slaves and concubines were the only women with whom a Muslim was allowed to have either premarital or extramarital sexual relations. While the number of wives was limited to four, the number of concubines remained limitless. A childless concubine could be sold, but rarely was.*® Once a concubine gave birth to a child by her master, she became in most [Islamic states an umm walad and could not be sold. On her master’s death, such a woman customarily was freed. All children born of legal concubinage were legitimate and usually inherited equally with children born in wedlock. The master had to acknowledge paternity, although in all legal traditions except the Hanafis, he was obliged to do so if the concubine was already umm walad.”’ Islam also encouraged men who were too

poor to acquire free wives to marry converted female slaves instead, although “such unions suffer only half the punishment for adultery reserved for formally free wives.”

Practice followed religious precept quite faithfully throughout the Islamic world, although there were occasional exceptions. (In Somalia, for example, the concubine was not usually freed after bearing a son, although she

was given an allowance.”’) The free concubine and her free sons are part and parcel of Islamic history and society. Many Islamic rulers have been the children of slave concubines, and the course of Islamic history has been decisively influenced by this pattern of manumission.”! Manumission by means of concubinage and marriage is by no means pecultar to the Islamic world. Typical of the preliterate world are the Sena of Mozambique, among whom “marriage between akporo [slaves] and free Sena were a form of institutionalized manumission.”°* Free Sena of both sexes could and did marry slaves. But even among preliterates, only a minority of peoples permitted free women to marry slaves; and among ad-

Manumission: Its Meaning and Modes = 229

vanced peoples with highly developed slave systems, the practice usually was strongly forbidden. Yet it would be a great mistake to assume that advanced premodern peoples always prohibited marriage between free women and slaves. The average male slave everywhere generally was unable

to marry a free woman because he could not afford to do so. Even where there was disdain for such marriages, exceptions were often made for prosperous slaves or those of powerful owners. In imperial Rome, where strong sanctions were imposed against free women marrying slaves, exceptions were made for the more powerful slaves of the emperor. The important point, however, is that marriage to a free woman did not necessarily bring with it free status for the male slave as it invariably did in the case of the female slave married to a free man. Women rarely confer status on their husbands, even in matrilineal societies where they determine the status of their children. To return to the more normal practice of manumission by means of marriage to a free man or concubinage with the master, we find that the practice was widespread throughout the world, in both precapitalist and modern times. China presents an interesting case. Contrary to Islamic practice, free women could also become concubines. What is more, it was possible to sell a free concubine, as long as her honorable or nonslave status was

made clear. Hence, it was even easier to identify the status of slave lover with that of free concubine. As in the Islamic world, the children of concubines often attained considerable status in the imperial hierarchy, sometimes becoming emperors themselves, and there are occasional cases of former concubines becoming empresses. What our analysis of the comparative data suggests is that, first, it is extremely unusual to find a slaveholding society in which freemen, especially masters, were prohibited from—or in practice, refrained from—cohabiting with female slaves with the inevitable result of producing children by these women. Indeed, I know of only one case in the entire annals of slaveholding societies in which female slaves were not sexually exploited and in which masters were strongly and effectively prohibited from cohabiting with their slaves. This is the extremely austere Gilyak of southeastern Siberia. Female slavery was an important part of their domestic economy up to the end of the nineteenth century, and possession of these human chattels was both an index of great wealth and a sign of prestige. The slave women, though economically and socially valuable for their masters, were viewed with disdain as persons, and a master who cohabited with one of them immediately lost status and incited great indignation in his community. This moral prohibition reinforced the domestic exploitation of the female slaves, for as Leopold von Schrenck, who studied the Amur group during the mid-nineteenth century explained, the sexual contempt in which the slave women were held meant that the wives of the master entertained no feelings of jealousy to-

230 Slavery as an Institutional Process ward them and welcomed them as workhorses within the household while they, the wives, performed the more pleasant handicraft tasks. Without doubt this sexual avoidance of slave women was largely due to racial scorn for the Ainus, from whom most of them were purchased.” Second, the sexual exploitation of slave women by their masters often resulted in ties of affection between them; and obviously, those ties were reinforced when the slave woman bore a child for the master. Many societies in addition to those advocating Islam automatically freed the concubine, especially after she had had a child. About a third of all non-Islamic so-

cieties fall into this category. In addition to the Sena, examples are the Nkundo, the Ibos, the ancient Vietnamese, the ancient Mesopotamians, Swedes certainly by the late thirteenth century (and possibly earlier), and the French Antilles up to about the end of the seventeenth century. Ancient Mesopotamia is of special interest. The Hammurabi code required the manumission of a female slave and her children on the death of the owner. It was customary for a wife who was sterile to give her husband a slave concubine in order to bear him children. Such assatum or concubines did not have the special slave tattoo and could not be sold, but their status was ambiguous; they remained the slave of the master’s wife while serving as

the master’s concubine. As one would expect, this was an emotionally charged situation that often led to tensions, especially after the concubine bore a child and became (inevitably) forward in her relations with her mistress. Apparently, prosperous and caring fathers gave such slave-maids at the time of their daughters’ marriages to ensure that their daughters were not divorced on the grounds of sterility. The biblical story of Hagar, given to Abraham by his sterile wife, Sarah, ended in jealousy and grief, and this may well have been the fate of many such triangular relationships. The ménage a trois hardly ever works, especially when a slave or servant is involved. C. R. Driver and J. C. Miles suggest that such Hagar-type concubinage occurred in the ancient Near East mainly where the wife was a priestess. Apparently in the late Babylonian period concubines remained slaves and could even be sold.”* In the second group of non-Islamic societies there was no legal requirement to free the concubine, but in practice it was usually done. Most nonIslamic societies fall into this group: the Mende, the Aboh, the Karebe of Africa, the peoples of ancient India and pharaonic Egypt, and the slaveholding societies of classical and medieval Europe. We have little evidence about concubinal relationships in classical Greece, although there are clear indications of its existence in the /liad and the Odyssey. The record of the Delphic manumissions in the second and first centuries B.c. suggests that free-

dom through concubinage with a master or other freeman was not uncommon. According to Keith Hopkins, we simply do not know where slave women procured the substantial sums demanded for their release in the Delphic manumissions; concubinage with a third party is a reasonable

Manumission: Its Meaning and Modes — 231

guess.” While we can only speculate on the instances of concubinal manumission in ancient Greece, ancient Athens must have been one of the rare cases where male slaves achieved freedom by this means. It is well established that there was an unusually high incidence of homosexuality among the slaveholding class, and it may have been that a substantial number of slaves bought for exclusively sexual purposes were males. Hence ties of sexual affection resulting in manumission may have applied as much to males as to females. The young boy slaves and youth freed in the will of the Peripatetics perhaps were cases in point. It is hard to see what other purpose they could have served in the philosophers’ households and for what other reason they should have been so favored in their masters’ wills.*°

Much less exotic was the Roman experience. Indeed, the traditional practice of manumitting concubines and their children was framed into law by Justinian, who provided that the concubine of a man who died without indicating her status in his will should be freed along with her children.”’ The practice of manumitting the concubine and her children continued throughout the Middle Ages and was also widely practiced in the slaveholding societies of the Americas, the Mascarene Islands of the Indian Ocean, and the Banda Islands at the southern end of the Moluccas. The view that there was a greater tendency toward concubinage in Latin America as opposed to the northwest European slaveholding societies is a myth, although it is true that manumission rates were higher in most of the Latin American

slave societies. Concubinal manumission was not uncommon in the English-speaking and French Caribbean slave societies. Indeed, genuine concubinage amounting to common-law marriage was found as frequently

in the British, French, and Dutch Caribbean as in other areas of the

Americas, including Spanish America. The Jamaican and French planters were sometimes as eager to manumit and endow their illegitimate offspring and their concubines as any Islamic master. Indeed, so much wealth was being passed on to the freed colored class that legislation had to be passed repeaiedly to limit the amount of money that freed colored offspring could inherit. The frequency with which the laws were reenacted attest to their ineffectiveness. It should be emphasized, however, that while a disproportionate number of freed persons were the concubines and children of their for-

mer masters, this group of societies differs radically from those of the ancient Near East and the Islamic states in that there was no automatic legal

manumission of such persons, and, more important, the vast majority of concubines and slave progeny did not receive their freedom. South Africa and the French Antilles during the second half of the seventeenth century did have laws or official edicts requiring manumission of the progeny of concubines, but recent works have demonstrated that while there were indeed a few such manumissions and even marriages between masters and their concubines, on the whole the laws were largely ineffectual.” There is, finally, a small group of slaveholding societies in which slave

232 Slavery as an Institutional Process women, while sexually exploited, were rarely freed. In the premodern world the overwhelming majority were kin-based matrilineal societies. Slave concubines and wives were highly desired in such societies, as a means of acquiring children over whom the father had complete control. Hence, such

concubines and their children tended to remain slaves. The Ashanti and Imbangala in Africa were two noteworthy examples. Other instances will be cited in the next chapter, when we discuss freedman status. In the modern world, South Africa during the eighteenth and early nineteenth centuries and the U.S. South during the eighteenth and nineteenth centuries are the most striking cases.

ADOPTION ,

From concubinal manumission we move to adoption as a mode of release from slavery. This is also an ancient and widespread mode, though by no means a universal one. In sharp contrast with concubinage, adoption as a means of manumission is conspicuous for its absence in the numerous Islamic slaveholding societies. The main reason, no doubt, is that Muhammad, while he did not abolish adoptions, significantly transformed the traditional custom, one that was basically similar to the fictive kin assimilation found all over the world. Muhammad himself recommended clientage as a more suitable form of bond between nonrelatives: “If ye know not their fathers, let them be as your brethren in religion and your clients.”°’ Subsequent Islamic law came to recognize only one form of filiation: what is known as igrdr or “acknowledgment,” which, while closely resembling adoption, has certain important differences. For example, a man cannot acknowledge another who is not at least twelve and a half years younger than he is; nor can he acknowledge a person whose parents are known. This automatically excludes the very group of slaves who are most likely to qualify for manumission—the houseborn. Favored slaves purchased from pagan lands, however, were not infrequently acknowledged, especially if the master had no heir. Outside of Islamic lands, adoption was practiced among the great majority of premodern peoples although it was rarely a very important form of manumission. The mode was very common among preliterate peoples. We have seen earlier that the slave was marginally reintegrated in many such societies as a junior member of the family. Thus, emancipation by adoption was simply an extension of the assimilative process begun during enslavement. There was usually, however, a definite point indicated by a rite of passage in which a slave ceased to be a slave and became a fully adopted member of his master’s family. Often the adoptive process took place over several generations, so that a descendant of a slave after a certain generation was automatically considered a full member of the community. Among the

Manumission: Its Meaning and Modes —_ 233

Katchin of highland Burma, for example, the mayan or slave had the status of an illegitimate son—between a rich man and a poor son-in-law. Absorption increased by blood the ties that strengthened over the generations, and when the mayan was to be manumitted, he was “ceremonially received into the clan of which he [was] a real blood member.” © It is striking that such automatic intergenerational adoption took place even where there was marked hostility toward the adoption of first-generation slaves. The group that best illustrates this pattern is the Gilyaks. We have already seen that first-generation slaves among them were viewed with total contempt. Masters, however, procured spouses for their slaves, and according to Lev Schternberg, the son of two slaves was considered “free to the neck,” that is, only his head was free. The grandchildren of slaves were said to be “free to the belt,” the great-grandchildren were “free to the feet,” and all descendants of slaves of the fourth generation were considered “pure.” They became k: khal, or kinsmen of their masters, and could sue anyone who called them slaves. “Pure,” however, seems to have been a relative term. Such manumitted persons, while pure in law, still were stigmatized and treated condescendingly.°! K. Nwachukwu-Ogedengbe, in his study of slavery among the nineteenth-century Aboh, an Igbo-speaking group of the lower Niger, brings out the difference between the initial fictive kin absorption of slaves and marginals and the status of the genuinely adopted freed slaves, and emphasizes the fact that such adoptions were highly unusual. He writes: “Although the slave was absorbed into the kinship system he was not included in the inheritance regulations governing true kinship relations. No property passed to him upon his master’s death except at the will of the master’s heirs. The only exception was an adopted slave ukodei on whose behalf the special ritual of igoya n’obii [to admit into the lineage through ritual sacrifice] had been performed. It must be emphasized that the ukodei was not a pseudokinsman but a kinsman in the fullest meaning of the term. The status of ukodei was, however, conferred only in exceptional cases—for example, in the absence of a natural heir who could sacrifice at the shrine of the deceased master” (emphasis added). Almost exactly the same observation could be made of the medieval Scandinavian communities. The child by a slave woman, if he was to be fully freed rather than having his descendants gradually absorbed intergenerationally, “had to go through a full ceremony of adoption into the family before he had equal rights of inheritance or received and paid a full share in atonement.”® Adoption may therefore have come about in two ways, by a gradual intergenerational process taking anywhere from two to five generations, or in a single stroke. The latter was usually attended by more ceremony, and the

assimilation of the individual and his descendants was apparently more complete than in the intergenerational case. Even so, in both instances it

234 Slavery as an Institutional Process would seem that the blemish of having a slave ancestor was never completely erased, even though the sanctions against any verbal abuse of the descendants may have been so strong that it was expressed only among intimates. Among a wide range of tribal peoples we find the adoption rite being repeated after long intervals in order to “cleanse” the ex-slave of all status pollution and to reaffirm the community’s acceptance of his changed status. Among many primitive Germanic groups the manumission of the slave had to be announced to the assembly of free persons twice in twenty years, after which no one could contest the change of status. And among the northwest coast Indians, when a free woman took the unusual step of marrying a slave (invariably a shotgun marriage), not only was the slave’s freedom bought and her husband formally adopted into the woman’s kin group, but ritual

cleansing ceremonies were conducted for the unborn fetus and were repeated later during various periods of the child’s lifetime. A more direct contractual type of adoption was practiced among most of

the literate and politically advanced premodern peoples. Although not mentioned in the Hammurabi code, release by adoption was one of the two most common forms of manumission in ancient Mesopotamia. According to

Isaac Mendelsohn, “Release by adoption was fundamentally a business transaction, a quid pro quo proposition. The manumitted slave entered into a sonship [or daughtership] relation with his former master. The relationship terminated with the death of the manumitter.” The release by adoption of a female slave was often accompanied by her marriage to a free man, in which case the couple was supposed to support the manumitter until his or her death. Adoption was rarely an act of generosity. In the Greco-Roman societies, adoption was somewhat unusual. Given the ethnically exclusive nature of Greek society, this is not surprising. The practice existed in early Roman society but was extremely unusual even during the late republican era. By the period of classical law, it had completely died out. Manumission by adoption was equally rare in the romanized areas of medieval Europe, and of course it was virtually nonexistent in the modern Americas. POLITICAL MANUMISSION

In one sense a form of adoption, political manumission occurs when the state or agent of the community (in the person of the chief, sultan, or ruler) adopts the former slave as a full-fledged member of the community with or without the consent of the owner. There were many reasons why the central authority or ruler of the community may have wished to manumit slaves, the most common being exceptional acts of valor on the part of the slave, usually in warfare. Typical was the old Norse law that “when common danger calls all (free and slave) to arms in the defense of the country the slave

Manumission: Its Meaning and Modes — 235

who succeeds in slaying an enemy in battle is free.”°’ As we shall see in Chapter 10, slaves were often freed by the state in order to make them eligible to become soldiers in societies where as slaves they were strictly forbidden to bear arms. Slaves who revealed treasonable acts on the part of their master or others also were frequently rewarded with their freedom, although this obviously was a dangerous practice if the charge failed to hold up. In ancient Greece the presentation of facts (as distinct from testimony) by slaves on such matters was one of the few occasions on which the slave could participate in the legal process. Where the state owned slaves, it could release them either on an individual basis or by a general pardon. In Greece, Rome, and in most other premodern societies ownership and release of such slaves presented few special problems. The situation was dif-

ferent in China, and to some extent Korea, where a large proportion of slaves in private hands were actually state owned, with their usufruct granted to favored officials. Whether the private masters of such slaves had the power to manumit them remained an unresolved legal problem in imperial China. More complicated were the periodic large-scale pardons of a substantial number of slaves, both privately and publicly owned, by Chinese emperors in occasional fits of magnanimity. From the continuing presence of slaves (and other evidence) it would seem that both the restrictions on the manumission of state-owned slaves by private masters and the manumission of privately and even publicly owned slaves in response to imperial decrees were frequently neglected.® Slaves were sometimes freed by the state when badly mistreated by their masters. This was not uncommon in some Islamic lands, since the Koran requires it, but it was also the case in a number of non-Islamic societies. Whatever the reason, manumission by the state was often the most complete method—and from the slave’s point of view, the one that granted him the fullest integration into the society. Among the Germanic peoples freedman status was often hemmed in by considerable civil disability. A major exception was the slave freed by the king for exceptional merit; he was immediately made a full member of the community and given arms. A similar situation prevailed among the Somalis, where freed persons usually were at a considerable disadvantage. However, where a person was manumitted by a decree of a sultan, he became a full member of the Somali community with none of the usual restrictions on his freedom. His natality was fully restored and he could, for example, marry a freeborn Somali woman.°’ In ancient Rome too, only the emperor could restore natality—that is, exercise the legal fiction that the former slave was born a freeman, whatever may have actually been the case, or that he had never been enslaved. If we are to accept the complex and subtle arguments of David Daube on the nature of early Roman manumission, manumissio censu (in which the slave was freed by being enrolled by the censor) was a distinctive form of

236 Slavery as an Institutional Process political manumission. Unlike other modes where the master released the slave and gave up what belonged to him, here the state selected and incorporated the slave as a citizen. Daube writes: “In manumissio censu, enrollment, incorporation by the State, came first, and liberty from the master was in strictness, only a consequence of that act. The master did not give up what had been his: he lost it, as a result of a political act.”°* Usually the master agreed to this act, but the censor had the right to register the slave against his

master’s will. While manumissio censu developed after other modes of manumission, it may have been the earliest mode of granting the freedman citizenship: the other modes merely granted release from slavery. Only much later did citizenship come automatically with the granting of manumission by Roman masters using all the different modes. When this happened—during the late republic—there was no longer any special need for manumissio censu, and by the period of the empire it had become obsolete. Buckland argues that its successor in the later imperial period was manumissio in sacrosantis ecclesiis, in that “it retains a trace of that element of public control which is dying out in the other forms.°” COLLUSIVE LITIGATION

The collusive mode of manumission was one of the earliest secular ways of circumventing the inalienability problem of release from slavery. It was a legal fiction, a form of collusive litigation that had much the same concept behind it in both Greece and Rome. In Athens it took the form of a simu-

lated trial in which the slave was tried for abandoning his master, and thereby his status as slave, with a predetermined verdict of acquittal. The acquittal was proof that the person was not a slave. It did not imply, however, that the slave, so freed, was a citizen. He became, instead, a metic.’° In Rome, where the procedure was known as manumissio vindicta, it took the

form of an adsertor libertatis claiming the slave to be a free man before a magistrate. The master, as in Greece, made no defense, and the slave was declared free. This mode, incidentally, was accompanied in ancient Rome by an unparalleled ritual: the master held the slave by one of his limbs, slapped his cheek, then turned him around.’! Buckland, like most other commentators, has expressed complete mystification at this extraordinary practice. Actually the practice of striking something or someone being alienated from its possessor, as a symbolic way of severing ties with it, is widespread. Mauss’s explanation is that something owned is felt to possess an element of one’s self in it and that therefore some symbolic means was necessary to break the bond when it was sold or given away. He found this to be true even of the France of his day, where numerous “customs show how it is necessary to detach the thing sold from the man who sells it, a thing may be slapped, a sheep may be whipped, and so on.” For Mauss such ritu-

Manumission: Its Meaning and Modes — 237

als were vestiges of the previous practice of the gift exchange, lingering in formal legal exchange. Indeed, what he writes of such vestiges in modern

France would hold even more for survivals in ancient law and society, barely removed from the natural economy in which prestations formed the dominant pattern of exchange: “The theme of the gift, of freedom and obligation in the gift, of generosity and self-interest in giving, reappear in our own society like the resurrection of a dominant motif long forgotten.”’” It is noteworthy that in Greece collusive litigation was the most common mode of manumission except for Delphi. It was, for example, the typical mode in Athens. In Rome it ranked second only to testamentary manumission. ”°

SACRAL MANUMISSION | The origins and development of sacral manumission are not at all clear, although the subject has received considerable scholarly attention. It may have existed in the ancient Near East, although the evidence is inconclusive. It was, however, the most popular form of manumission in Delphi and thus is firmly associated with the Greek world. It is important to emphasize that

sacral manumission was not the most popular form of manumission throughout Greece. It is merely the one about which we happen to have substantial evidence. It was not, so far as we know, practiced at any of the major Greek cities such as Athens, Corinth, or Thebes, where collusive litigation and testamentary manumission seem to have been the standard modes of release.’* According to Bomer, there were two kinds of sacral manumission: fiduclary consecration and fictive sale to a god. In consecration, the older of the

two forms, the manumitter initiated the process in the hope that the god

him.’° |

would free the slave; in fictive sale to the god, the slave initiated the process (practically, though not legally) under the condition that the god would free Though not the first to do so, the Delphic priests articulated most clearly what was entailed by manumission in terms of four basic freedoms: “[lega]] status, personal inviolability, the right to work as one pleased, and the privilege of going wherever one wished.””° There is no source in Delphi older than 201 B.c. BOmer rejects F. Sokolowski’s thesis that sacral manumission originated in the practice of the slave’s seeking asylum in the temple.’’ Another hypothesis is that sacral manumission may have developed from the genuine sale of slaves to the temple—in other words, temple slavery. Bomer expresses some doubt about this view, on the grounds that there were none of the expected transition forms between temple slavery and the fictive process of sacral manumission.’* To argue from silence is always dangerous, but the comparative data support Bémer’s skepticism. It is striking that wher-

238 Slavery as an Institutional Process ever we find temple slavery, there is always a strong prohibition on the sale of such slaves. This was true, for example, of the temple slaves of ancient Mesopotamia, the sirgu, who became a hereditary caste of slaves and who were the only group of slaves barred from manumission.” It is significant that the slaves of the early and medieval churches were usually the last to be freed in the transition from slavery to serfdom.*’ In Burma we find that the temple slaves were a despised hereditary caste and could not be manumit-

ted, even by the ruler.*' And to cite a final case, the Osu, or cult slaves among the Ibos, were the one group who could never be assimilated. Even today descendants of Osu slaves, even upper-class Osus, suffer the stigma of their ancestry—unlike the descendants of other slaves.*’ Clearly there is something extraordinary about the status of the temple slave.” I am therefore fully inclined to accept BOmer’s thesis that sacral manumission developed not before, but after secular manumission in Greece, and I certainly agree with his rejection of the neoevolutionary assumption that the secular always follows the sacred form in the development of institutions. Sacral manumission developed as a substitute for adequate legal processes, as a way of giving sanction and ceremony to a purely secular legal act. The “guarantees were stronger when the participants considered the gods involved, while the essence of the operation was legal.”** The best support for this is the fact that when the authority of the state was strong (as in Athens, Corinth, and Thebes) there was no need for, and no surviving evidence of, sacral manumission. Hence, sacral manumission was essentially a consequence—and a late one—of small communities with poorly developed political and legal traditions, or of the decline of the polis.*° The Delphic manumissions are extremely revealing about such practices in the ancient world. As Hopkins observed, they constitute the “hardest” set of data we have on almost any subject in antiquity. They reveal not only the “degraded” condition of slaves, but the fact that slaves paid dearly for their freedom.”° The average price of 400 drachmas at the end of the third century was the equivalent of what it would cost “to feed a poor peasant family for over three years.” Freedom usually had to be bought in installments. What Slaves got in return was often of dubious material value. The down-payment terms (paramoné )}—which we shall come to later—made freedom often little

more than an “illusion,” according to Hopkins. What is more, the prices were constantly rising during the last two centuries before Christ.*’ THE FORMAL CONTRACTUAL MODE

We come, finally, to purely contractual manumission. Most forms of manumission, of course, had a contractual element. We have seen that Babylonian adoptions were largely business documents. Even so, they were a legal fiction, as was manumissio vindicta in Rome. We are thinking here of what

Manumission: Its Meaning and Modes — 239

Weber called legal-rational contracts, those that completely ignore the need

to circumvent the conceptual problem posed by manumission. Buckland suggested that Roman manumissio testamento rapidly developed into just such a contract. While it was a highly developed legal contract, it nonetheless retained “nonrational” symbolic elements. Formal contractual manumission really grew out of various types of informal manumission that made their appearance during republican times. Two such forms were (1) a simple verbal declaration by the master that the Slave was free, and (2) the writing of a letter of manumission before witnesses. Neither was valid during the republic.** Justinian formalized these methods, giving some of them the same validity as manumission vindicta. Among the most important of the new, formal methods were per epistolam, in which a letter of manumission by the master was witnessed by free persons; inter amicos, similar to the earlier version except that it was also signed by a magistrate; and simply destroying the slave’s papers in the presence of five witnesses.” They became more popular and more fully developed during the late ancient world and the Middle Ages, though perhaps were never the most important modes until rather late. In thirteenth-century Spain, for example, manumission by will remained extremely popular,” over the years manumission by church authorities gradually became the dominant mode. It is only in the modern world that the legal-rational mode became dominant. The situation was different in Islamic lands. We have already seen that

testamentary and concubinal manumissions were the most common, but from Muhammad’s day on, purely contractual manumission was provided for and encouraged. The Islamic practice, called the kitab in the Koran, resembled the Greek paramoné system: the slave paid for his freedom in equally spaced installments. Most Islamic authorities insisted on the installment mode of payment, only the Hanafis sanctioning a single payment. The master cculd not sell the slave during the period of his payment. When the mukataba, as such a slave was called, completed his payment, he was free, and a rebate was usually given him.”! I BEGAN this chapter by observing that the release from slavery posed three kinds of problems: the conceptual issue of how to define the transaction in

meaningful terms; the cultural problem of giving symbolic and ritual expression to the transaction as well as customary form; and the more social problem of creating a new status for the freed slave. Having explored the first two of these areas, I turn in the next chapter to the third.

The

Status of Freed Persons |

| |

THE ACT OF MANUMISSION Creates not just a new person

and a new life, but a new status—into which we must now inquire. The freedman must establish two kinds of relationships: with his former master, and with society at large (more particularly, with free men other than his former master). The two are closely related. Indeed, the single most important factor determining the condition of the freedman in the society at large will be the nature of his relationship with his former master.

The Freedman and the Ex-Master There is a remarkable uniformity among the slaveholding peoples of the world with regard to this relationship. Almost universally the former master has established a strong patron-client bond with his freedman. In most societies this bond has been sanctioned by law. Actually, the intrasocietal variation in the nature and strength of the patron-client relationship tends to be greater than the intersocietal variation; that is, in any given society the nature of the dependence of the freedman upon his former master was affected by economic matters, such as the price the slave paid for his freedom (in the event a price was paid) and the terms of the payment, as well as the sex of the slave (women usually being more bound than men), his or her occupation, and the nature of the preexisting master-slave relationship. 240

The Status of Freed Persons 241 When we look across societies, the similarities were impressive. Everywhere the freedman was expected to be grateful for the master’s generosity in freeing him, however much he may have paid. This followed naturally from the universal conception of manumission as a gift from the master. In Rome a freedman could be charged with the crime of ingratitude and, if guilty, could be reenslaved. Much the same pattern existed throughout medieval Europe. Other societies were less legalistic about the matter but no less demanding. Everywhere in the premodern world the freedman had to honor his former master, and everywhere certain social obligations were expected of him. “Patronage belongs to the emancipator,” goes the Islamic saying. Indeed, one must see the relationship between ex-slave and ex-master as something quite distinct from the normal patron-client relationship, entered into freely and voluntarily by nonslave persons. The relationship between ex-slave and ex-master was always stronger and always carried with it a certain involuntary quality that was quite distinctive. It cannot be viewed in isolation from the relationship it replaced. For this reason I propose to use the Arabic term wala to distinguish this relationship from client-

ship among free persons.'

In many lineage-based societies the freedman was assimilated, though in an inferior capacity, into his master’s lineage or clan or family. Almost invariably in these cases the freedman continued to function within the same economic context as previously. He had little choice. Land was corporately owned, and access to it was determined by kin ties. This financial dependence was not usually a hardship, since enslavement in such societies was not primarily economic in origin. Freedman status with respect to the master tended to have mainly social and psychological implications. The freedman became legally competent, could sue and be sued (usually not his former master). He could now own property and had custodial powers over his or her children. The ex-slave could marry without the ex-master’s consent, and the range of potential spouses was wider—though rarely as large as it was to persons who were never enslaved. In Islamic societies the freedman and his descendants established a hereditary bond of kinship with the former master and his descendants. Both

patron and client were referred to as mawala in the wala relationship. In China, filial piety and respect were expected from the freedman. In the ancient Near East, the freedman reputedly was not truly free until his master’s death, so strong was the bond. In the advanced slave systems of Greece and

Rome, ties between patron and client were equally strong, sometimes assimilated to the relationship between parent and child. Many Greek freedmen spent the rest of their lives paying off the remainder of the mortgage for their freedom, the result being that they remained firmly attached to their ex-masters. Freed persons were frequently required to obey and respect their former owners and to serve them for the rest of their lives. Sometimes these obligations were transferred to the ex-master’s heirs. In Athens if

242 Slavery as an Institutional Process a patron successfully sued his freedman for disobedience, the latter was returned to slavery; however, if the patron lost, the freedman became absolutely free. Between 340 and 320 B.c. there were approximately fifty such cases each year.” There were marked variations in the kind and strength of the wala relationship. In the relatively complex urban economy of the Greek city-states,

opportunities existed for some slaves who so desired to break the bonds completely, either by moving to another part of the country, to another part of the city, or entirely away from the country. The tremendous range of occupations of slaves meant that some slaves were better able to take advantage of their freedom than others. Rural slaves were not only far less likely to be manumitted than urban ones, but in the unlikely event that they were, their dependency ties remained much stronger because they were less capable of surviving on their own.” Before Islam, nowhere was the wala relationship more elaborately prescribed both in law and in practice than in ancient Rome during late republican and imperial times.* There were three kinds of claims that the patron could make on his freedman. First, there was the obsequium. This basically meant the showing of proper reverence and gratitude to the patron and his kinsmen. It is not clear whether this attitude was legally enforceable during republican times. Susan Treggiari argues that it could have been legally enforced only if it had been stipulated at the time the manumission contract was drawn up. Otherwise the claim had moral sanction only, albeit this was powerful. During imperial times obsequium increasingly became a legally enforceable claim of the patron. The second and more practically significant claim of the patron was the

operae. This was the obligation of the freedman to work for the patron, which “sprang, not from the status of libertus, but from an oath which the freedman took after manumission.”” Almost all masters insisted on such an

oath. In classical law it was legally enforceable and automatic, but even from republican times it was strongly established “that operae was naturally owed to the patron in gratitude for the supreme gift of freedom.” The operae

referred to units of days-work, hence it was possible to specify a certain number of operae per year, or even the total number of operae due. The freedman was expected to perform work for which he was qualified, except that he was not required to perform tasks that endangered his life even if, as a Slave, he had been trained to do so. Operae claims could be transferred by the patron to another person and were passed on to the patron’s heirs. The third kind of claim the patron had on his freedman was the right to

half, and 1n some cases all, of the freedman’s estate on his death. These claims were also inheritable by the patron’s heirs.

The patron’s claims on his freedwomen—his libertae—were even stronger than those on his liberti since, in addition, they were under his tu-

The Status of Freed Persons 243 tela. What, in practical terms, this meant was that he or his heirs always inherited the entire estate of his libertae. The wala relationship was supposed to be reciprocal. But since the obligations of the freedperson were conceived as expressions of gratitude for the gift of freedom—however great the material rewards that may have accrued to the ex-master—it is understandable that the patron’s obligations were few and vaguely defined: he was to protect and aid his freedpersons as best he could.

In view of the symbolism of gift exchange inherent in the act of manumission, it is not surprising to learn that in spite of its enormous economic significance for both patron and freedman, the Roman wala relationship ul-

timately rested on moral rather than legal force. Reinforcing the law, at times even functioning without it, was the powerful Roman sentiment of fides (trust, faith, honor, loyalty, allegiance). Treggiari concludes her analysis with the observation “that the whole structure of obligations and rights between patron and freedman rested on the moral concept of fides and that the law sought to strike a balance between conflicting interests.” As in the Greek city-states there were considerable individual and regional variations in the kinds of balance struck between patrons and freedmen. A few talented and lucky ex-slaves were able to buy themselves completely out of the indignities of spiritual and economic dependence. Others, the great majority, had no choice but to prolong the relationship for the rest of their lives and, what is more, pass it on to their children—gaining, in return, the gift of freedom, a gift that may have been materially meaningless but nonetheless meant moral worth, belonging, and self-respect. Furthermore, in the isolated case of Rome, it also meant, if the master was a Roman citizen, the remarkable gift of citizenship, a privilege that has no parallel in

the history of slavery. |

Roman law greatly influenced most of the slaveholding systems of medieval continental Europe, so it is no surprise to find a replication of the patterns just described in Visigothic and later in Christian Spain, as well as in France, medieval Italy, and the large-scale slave systems of the Mediterranean islands during late medieval and early modern times.’ There were, to be sure, variations based on peculiar local customs or pre-Roman law, but the general pattern was to adopt Roman practice and to codify in law what earlier Roman practice had left principally to moral suasion. The northern Europeans require special comment, in view of the lateness of the Roman legal influence. Here, as elsewhere, the wala relationship was strongly enforced in both traditional law and custom. As Thomas Lindkvist

points out in his study of the /andbor and related classes in the Nordic countries during the early Middle Ages, the freedman remained under a strong bond of personal dependency throughout the Scandinavian lands, and his relation to the ex-master was sharply differentiated from that of

244 ~~ Slavery as an Institutional Process

other semifree persons to their lords. The relationship was inherited by the heirs of the patron and continued for as much as two generations among the descendants of the freedman. The control of the patron seems to have been even stronger than among the central and southern Europeans: for example, he continued to have a say in the freedman’s marriage. As in other parts of the world, there were some variations in the amount of control. Freedmen who could afford and were permitted the ritual liberation beer party were under no more obligation than other full members of the patron’s family, which is not to say that a dependency relationship did not persist. However, intrasocietal variations in the degree of dependence were much smaller in these essentially rural Nordic countries; almost all freedmen joined the economic, if not the social, ranks of tenant farmers.° Intersocietal variations in the wala relationship were much greater than intrasocietal. At one extreme, the laws of Jutland and Sjaelland suggest that freedmen were under no further obligation to their former masters, and this is explicitly stated in the Sjaelland laws after 1215. At the other extreme were the Norwegians, among whom the freedman remained under the tight control of his former master: A man in this situation had no freedom of movement, he owed his patron certain dues in labor for one year, he had to consult him on any business, 1ncluding marriage, and he shared any atonement for injury with him. If he conspired against his former master or joined his enemies or took part in a law suit against him or “spoke to him as if on an equal footing,” then he forefeited his property and returned to servitude. On the other hand, the patron took responsibility for the freed slave’s maintenance and gave him general support.°

In certain parts of Norway the wala continued for four generations, only those descendants born of the fifth generation being free of the dependency. We come finally to the modern world. Here, as everywhere else, ties of economic dependency with the ex-master were the norm. However, there were important differences in the degree to which the wala was institutionalized and legally enforceable. In only a small minority of American societies was there a universal application of a formalized wala: among these were the slave societies of the Dutch Antilles. According to Harry Hoetink, “freed slaves and their offspring were obliged to show all honor, respect, and reverence to their former master, his wife, children and their descendants. Offences against his former master could result in the freedman’s reversion

to slavery.”'® In the Spanish Americas and colonial Brazil the wala was never legally formalized. However, masters did manumit slaves on condition that the freedmen perform the equivalent of the Roman operae: “Slaves were freed, for example, on condition that they continue to work for their former masters for a certain period of time each day; or a black owned by a partnership might become, say, one-third free upon manumission by one of the partners, in which case he would divide his time between his own occu-

The Status of Freed Persons 245 pations and continuing service to the partnership.”'' We also find conditional manumissions that were highly reminiscent of the contractual obligations imposed on freedmen in medieval Scandinavia. “In rural areas, some masters freed large numbers of blacks but made sure that they became tenants. The former were thereby freed of the costs of slavery, assured of a fixed annual rent from the lands involved, and at the same time kept a pool of

labor to draw upon at harvest time.” !” ,

The situation was much the same in colonial Brazil, where slaves were

not only frequently manumitted on condition that they perform certain tasks but an individual could be and was reenslaved “as the law decrees for repaying by ingratitude the favor of having been granted his freedom.”'” Nevertheless, even in Latin America conditional manumission in the strict legal sense of partial freedom until clearly specified conditions were met constituted a minority of all cases (in some areas a substantial minority). In Buenos Aires between 1776 and 1817, such manumissions made up 10.9 percent of all cases; in Bahia between 1813 and 1853 they constituted 22.5 percent; in Paraty, Bahia, between 1789 and 1822 they made up 42.5 percent; in Lima between 1580 and 1650 the figure was 18.4 percent; and in Mexico City between 1580 and 1650 it was 24.3 percent.'* When one adds to these percentages those of slaves manumitted gratis, the total in all of these societies except Buenos Aires ranges between 52 percent (Lima) and 66 percent (Paraty). And when it is recalled that slaves manumitted gratis were nearly always under powerful moral pressure, not to mention economic and political pressure, to display gratitude and respect to their manumitters, it 1s seen that the wala existed in practical terms for the majority of freedmen. As for the remainder, what Stuart B. Schwartz observes of all freedmen in colonial Bahia would have held true of freedmen throughout Latin America, that their manumission was “ultimately conditional in that a /iberto was always subject to reenslavement”; even if such laws were rarely put into practice, “the very threat of enforcement may have been enough to produce the

desired result of social control.”!° |

The wala relationship was least formalized in the more highly capitalistic slave systems of South Africa, the British Caribbean, and the U.S. South. In South Africa during the less vicious period of the late seventeenth century, “quite a number were freed unconditionally,” according to A. J. Béeseken, although many freedmen were subject to specified operae and other conditions, some of them quite extraordinary: one Paul de Kock, for example, was

freed on condition that during consecutive periods totaling more than five years he serve two men, both of whom lived in Batavia, several thousand miles away from South Africa.'® In the eighteenth century the South African pattern came to resemble its southern U.S. counterpart. There were very few manumissions, but the great majority (84 percent) were wholly unconditional: “freed slaves did not become thinly disguised indentured servants.”'’ What is most revealing is the

246 Slavery as an Institutional Process difference in the experience of privately owned slaves and of the semipublicly owned slaves of the Dutch East India Company. The company manumitted its slaves at twelve times the rate of the private owners but “imposed the harshest conditions on its freed slaves.”'* This intrasocietal difference in the experience of slaves in South Africa highlights an important intersocietal pattern about which I shall have more to say in the next chapter, namely, the higher the rate of manumission, the stronger and more formalized the wala, and vice versa. The fact that the wala was not legally enforceable in most areas of the New World meant that those freedmen who wished to break the ties could do so. A few did, but only a minority; while many more would have liked to,

‘they were hampered by the same set of constraints that applied to their Greco-Roman counterparts: the purchase of their freedom may have left them penniless or indebted; their lack of urban skill may have confined them to rural areas, where their lack of land meant that they could only turn to the ex-master for help; even where they had a skill, they may have been

confined by their white competitors to the lowest-paying jobs, so-called “nigger work”; and, perhaps most important of all, affective ties to kinsmen and other loved ones still in slavery often dictated that the most expedient course of action was to maintain ties of dependency with the ex-master. A factor that was of special importance—though not unique—to South Africa

and the U.S. South was racist hostility on the part of free whites to the mixed-race or black freedman; the ex-master was often the only source of protection in a society whose laws were not only unsympathetic, but whose minimal protections were not enforced. Ira Berlin’s study demonstrates that while “free Negroes often went out of their way to break the bonds of dependence,” very few in fact succeeded in doing so.'” At the same time it was rare to find legally enforceable wala obligations in the United States or in the non-Dutch Caribbean. In South Carolina during the 1830s freed slaves were required to have legal guardians from the white population; these need not necessarily have been the former master. In any case, the purpose of the law was not to strengthen the per-

sonal bond between freedmen and former masters, but to police the former.” In the United States, as in other parts of the New World, manumission was highly selective, favoring those slaves who were most likely to want to

maintain the ties of dependence.*' In the urban areas of the South, especially the older parts of the antebellum South, most of those freed were women with sexual ties to the master; in the rural areas most of those freed were men, but it was precisely in these agricultural areas that freedmen were most at the economic mercy of the ex-master. Berlin chronicles the desperate economic marginality of the male freedmen, most of whom remained in farming, in a “vicious cycle of debt and de facto servitude.” Certain generalizations may be made with regard to the freedman/ex-

The Status of Freed Persons 247 master relationship as it existed in all societies. Masters rarely lost much in tangible terms—economic or political—by manumission and usually gained a great deal. Invariably the ties of dependency continued, usually fully institutionalized in the wala relationship throughout the premodern world, but in only a minority of cases in modern slave systems. Whatever the primary reasons for original enslavement, these continued to be served after manumission. If the main reason had been economic, economic ties persisted; if the main motive had been political, such ties were reinforced; if the primary motive had originally been sexual, invariably the former master continued to enjoy sexual satisfaction. If, as in most primitive societies, slaves had been mainly prestige goods, the freedmen joined the household of the ex-master and further enhanced his dignity and honor.”

| The Freedman and the Freeborn The status of the freedman in the community at large is the second major area to be considered. In a comparison across societies it is useful to distinguish between the political-legal status of the freedman and what may be called his prestige ranking. By the latter is meant the respect with which the freedman was viewed—the degree to which he was accepted as an equal who fully belonged to the community. Full political-legal capacity does not necessarily imply full social acceptance. Alternatively, there are a few cases where the freedman was fully accepted in prestige terms, yet did not achieve full legal and political capacity—this usually occurred where the ex-slave was a native-born person who fell into slavery for political or military reasons. A classic example is the Roman captured by the enemy who, after being ransomed, was freed postliminium but was subject to certain limitations with respect to his ransomer. Let us begin with the problem of prestige ranking. Nominally granted almost complete equality, politically and legally, with “free” persons, freedmen nonetheless remained stigmatized. Even among people such as the Sena, who went as far as possible in incorporating the manumitted slave, freedmen were still “treated condescendingly by junior kinsmen,” were obliged to perform the most unpleasant tasks, and were first to be sold if the family faced starvation in a time of economic crisis.“ The stigma of former slavery meant that the freedman was rarely perceived as an equal. Only time could blot out the memory of the debased condition he experienced as a slave. Hence, full freedom came only to his descendants. How long this took varied from one society to the next. In well over 80 percent of all significant slaveholding societies freedmen suffered some civil disability. Honored with the nominal status of citizenship, in practice they remained second-class citizens. In almost all societies ex-slaves were barred from the most important leadership roles in the com-

248 Slavery as an Institutional Process munity. (We exclude here, of course, the special case of palace slaves and freedmen, to be considered in Chapter 11.) Occasionally an ex-slave became a minor chief, as among the Mende, but such cases were always considered exceptional by the people themselves and could sometimes lead to trouble. The Duala of the Cameroon provide an instructive illustration. During the nineteenth century a captive married the daughter of a chief and produced a ruling family of one of the major Duala towns, Deido, but as Ralph A. Austen tells us, “The subsequent history of Deido is marked by particularly severe conflict with other Duala towns, culminating in the unprecedented execution of the chief, Charley Dido.”*° Even where freedmen were relied on to fill executive and administrative roles, it is important to recognize that the great majority of freedmen were excluded from such positions. Igor Kopytoff and Suzanne Miers make the important point that a range of slave statuses did not necessarily reflect significant mobility. Some slaves may have been acquired specifically for official tasks, others to work as laborers—and the latter may not have had the slightest chance of rising to the ranks of the former. The same holds for freedmen. In imperial Rome it was almost impossible for a freedman of a private master (the vast majority of all freedmen) to rise to the position of an imperial freedman, since these were recruited from within the ordo of the familia Caesaris. This was even more true in the Islamic caliphate and the Ottoman empire, where slaves were recruited for military and administrative roles from specific areas with very ascriptive criteria on race and ethnicity. A despised African Zandj who against all odds managed to win his free-

dom in ninth-century Iraq stood no chance of achieving the military or executive rank of the Turkish Mamluk freedman. The janissaries of the Ottoman empire were all white, the children of Christian subjects. I have repeatedly emphasized that slavery was not a static institution. From the moment the slave entered his status, changes began to take place in his relations with his master and with the rest of the community. Kopytoff and Miers neatly sum up this process by applying the common sociological

distinction between intergenerational and intragenerational mobility to three dimensions of the slave’s relationship: the slave’s legal status, his affective marginality, and his worldly success.”° They observe that changes along these three dimensions may occur during the lifetime of a single slave, that

is, intragenerational or what they call “lifetime mobility,” which must be distinguished from “changes that his offspring and descendants will experience, that is, intergenerational mobility.” Then they add: “The rather obvious distinction must be kept in mind because such statements as ‘the slave becomes integrated into the lineage in several generations’ have sometimes been taken as showing the flexible and benign nature of a slave system. It

should be remembered that intergenerational flexibility can coexist with rigid statuses into which each generation may be frozen.””’ These points are well taken and are fully supported by the comparative data.

The Status of Freed Persons 249 However, we come upon another ambiguity in the anthropological literature on slavery. Emphasis on the fluidity and intergenerational nature of slave status tends to give the impression that the freedman, once manumitted, was fully integrated, and this was far from being the case. In Charles V. Monteil’s discussion of slavery among the Bambara,”* for instance, he tells us that the status of slaves was not fixed, but generally passed through three stages. The children of slaves “born in the house” (wolo-so-u) had a privileged position—by which he means privileged vis-a-vis other slaves, although this is only implied. The third generation, he adds, “‘were ipso facto freed (dyongoron-u),” and he adds, “They have played an important role in native society.” Maybe so, but were they really equal to freeborn Bambara? Significantly, Monteil slips and refers to this third generation as “another category of slaves.” The very fact that the freedman belonged to a specially designated category of persons—dyongoro—suggests that he was indeed different, and we can deduce from the marked sensitivity to status among the Bambara (fully documented by Monteil) just how insecure must have been the place of the freedman. What was true of the Bambara was true of virtually all slave-owning societies, and what it amounts to is this: freedman Status was not an end to the process of marginalization but merely the end of the beginning—the end of one phase, slavery, which itself had several stages. Freedman status began a new phase: the ex-slave was still a marginal, but the process was now moving toward demarginalization socially, and disalienation in personal terms. The new phase may itself have taken several generations, although as with slavery, for a fortunate few the process may nave been short-circuited and the freedman immediately declared free. This was true, for example, of Somali slaves freed by the sultan as a reward for exceptional acts, and true of those Roman slaves who by imperial edict were granted the status of ingenuus (locally born free persons). By their very nature such cases were highly unusual. As a marginal person the freedman continued to be viewed as something of an anomaly and, like all persons in transitional states, was regarded as potentially dangerous. The community took an active interest in him not only for economic reasons, to ensure that he did not become a public burden, but also for social and symbolic reasons. We have already seen how in western Norway during the Middle Ages the ties of dependency with the master lasted for four generations. Paralleling this dependency was the contempt in which the freedman and his descendants were held for carrying the lingering stain of slave ancestry. The image of the thrall as nasty, ugly, foul, stupid, cowardly, and inferior had

been racist pure and simple; in exactly the same way, the attitude to the freedman was racist until the fifth generation, when at last “the stain” was removed and the descendant became “pure.” The New World slave societies differed from this situation only in degree. In the French slave colony of Saint Domingue during the second half

250 Slavery as an Institutional Process | of the eighteenth century, when hostility to the freedmen was especially strong, a book that had the official blessing of the French government declared, “Interest and security demand that we overwhelm the black race with so much disdain that whoever descends from it until the sixth generation shall be covered by an indelible stain.”*’ It is very tempting, on the basis of passages such as these, to hasten to the conclusion that New World slavery was unique in the additional burden of racism that slaves and freedmen bore. But this identification of a given group with slavery was as old as slavery itself; freedmen in all parts of the world suffered this perception of

them as being stained by both slavery and the group with which they were identified. Han freedmen among the Lolos were stained with descent from Chinese blood. Among the Ashanti the freedman who was of foreign ancestry was stained with both his slavery and his northernness. In the medieval Muslim world “Zandj” meant slave as well as black, and both “stained” the freedman and his descendants. Throughout medieval Europe “slave” and “Slav” became so indistinguishable that “Slav” came to mean “slave,” a linguistic fate that did not befall the words “Negro” or “black” in any of the European languages (even if there has been a sociological identification of blacks with slave status). In spite of these worldwide uniformities there have been variations between slave societies in the kind and pace of social reception of the freedman. What has determined such variations? Without doubt one of the most important faetors has been the degree of institutionalization of the wala relationship, the extent to which the relationship has been formalized and given legal-cultural sanction. Among the other variables that interact to determine political-legal status and prestige of the freedman the most important are race, the social formation of the community, and the demographic composition of the population—especially the sex ratio of the master class and the proportion of the total population who are slaves. The mode of manumission, itself closely related to these factors, operates independently in determining the freedman’s acceptance. As a result of the joint operation of these variables, I find the slaveholding societies of the world to be grouped into six types. _ (1) In the first group of societies, the freedman and his dependents retained close ties of dependency with the former master and his family for generations—indeed, in perpetuity. Over time the ex-slave was physically absorbed into the former master’s family. While there were always exceptions, freedmen’s descendants usually constituted the “poor relations” of the family. The degree and pace of absorption depended largely on whether the original ex-slave was a native who had fallen into slavery or an ethnic outsider. Race was not a factor, since both groups belonged to the same race. The freedmen and their descendants were legally full-fledged members of the community by virtue of being members of the former master’s family.

The Status of Freed Persons 251 There was usually little or no economic exploitation of the freedman or his descendants, given the fact that these societies were usually small-scale subsistence systems with little class development. I include here the great majority of lineage-based and other preliterate societies that kept slaves. There is, however, an important subdivision of this group. The freedman and his descendants were more deeply absorbed and the stigma more rapidly erased in patrilineal societies than in matrilineal ones. In matrilineal societies the vast majority of freedmen and their descendants could be easily recognized by virtue of the patrilineal origins of their clan name. If a slave concubine was the ancestral freedperson, her descendants had to trace their ancestry back to her master, who would have been the ancestral male. The major exception was the rare case in which the freedman or woman was originally a native free person who had fallen into slavery and was later restored to freedom, in which case he or she would have simply rejoined the matrilineal group. Among the Yao and Ashanti the stigma of slavery persisted for many generations for this reason. In most kin-based societies, however, the freedman and his descendants were physically absorbed in two or three generations. And in some, such as the Ila, almost all stigma was gone by the second

generation. |

(2) In a second group of societies, the fate of the freedman is determined mainly by the interaction with the wala relationship of gender and the mode of manumission. The majority of freed persons were women who were absorbed as concubines or wives. Their children were wholly absorbed into the family of the master, and the stigma of slavery disappeared within a generation or two. Male freedmen and their spouses and descendants experienced a separate fate. Ties of dependency, often economic and political, remained strong and perpetual. The freedman and his descendants become a distinct status group intermediate between slaves and free men, sometimes living in separate areas, sometimes maintaining strong economic relations with the former master’s family and close physical proximity to them. Although such freedmen may have been formally considered citizens of their society, they

continued to suffer the disabilities that all dependent groups experience: they were culturally assimilated but socially excluded. Race and sex operated jointly to produce two opposing effects. Where, as was usually the case here, the slaves were of a different racial group, female

slaves who became concubines were freed along with their children and were fully absorbed by the master class. Concubinage was legal and children of these unions inherited equally with other children. Male freed persons, however, were racially excluded. Barred from marrying women of the exmaster’s race, they tended to take their wives from freedwomen of their own

race, often buying them out of bondage or being given them for this purpose. The strong wala relationship established bonds of solidarity with the dominant race. There was no sense of solidarity with the slave group. Ironi-

252 Slavery as an Institutional Process cally, there often was strong prohibition against marriage or concubinage between males of the master class and freeborn women who were children of this freedman caste, even though there was a high rate of concubinage and even formal marriage with females of the slave class. The “estate” nature of such freedmen standings was usually reinforced by racial or somatic differ-

ences, although the genetic absorption of slave concubines tended toward some degree of somatic convergence. Reflecting this convergence was a marked sensitivity to somatic differences among all groups. This second group subsumes all the Islamic slaveholding societies (including those of the Sudan and the Sahel) with the exception of the racially homogeneous black African communities who advocate Islam. (3) In the third group of societies, there was little or no perceived racial difference between masters and slaves, although strong ethnic differences may have existed. The wala relationship was highly formalized in law. The freedman was made a citizen and had full legal capacity with respect to persons other than his ex-master and the latter’s family. There was a highly centralized social formation and the freedman was given citizenship, even though he was barred from certain of the highest and most prestigious positions. The freedman experienced some stigma; but this varied with his skill, education, and wealth, and whether he lived in a rural or urban area. The stigma was completely gone after two generations. Rome was the classic instance of this kind of society. The assimilation of

freedmen was particularly impressive in view of the fact that almost all slaves were of foreign ancestry. But we should be careful not to exaggerate. In theory the liberti of a Roman citizen were citizens. But as Treggiari has

shown,’ custom, law, and prejudice together ensured that they were second-class citizens. They could not stand for office in Rome and were usually prevented from holding offices in other Italian towns; they were apparently excluded by custom from magistracies elsewhere. The sons of freedmen suffered few if any of these civil disabilities. They became senators, held magistracies, and were admitted to the ranks of the equestrian offices. Nonetheless, the stigma of slavery still was suffered by the sons of freedmen, as the frequently cited case of the poet Horace attests.”! China, Korea, and Vietnam also fell into this group. The freedman was

restored to full citizenship within his lifetime, although the stigma took longer to be erased (about three generations). Chinese history is full of successful descendants of slaves who were slandered by their half-relatives.

Typical of these was Ts’ui Tao-ku, the son of a former slave during the period of the Northern dynasties (A.D. 386-618), who was so ill used by his half-brothers that his father gave him some money and packed him off to government service in the south. Ts’ui made good, and after a successful career returned home and, in triumph, held a banquet for the local officials. His cruel half-brothers would not be reconciled and insulted him by forcing his mother, now a freedwoman, to serve the dinner at the banquet.”

The Status of Freed Persons 253 The ancient Near Eastern societies must also be included in this category. In Mesopotamia and in pharaonic Egypt the freedman was made a “son of the city” or a “freedman of the land of Pharaoh,” both of which are taken to mean citizenship; this was accompanied by a wala relationship that

lasted until the master’s death. Jacob Rabinowitz’ finds the parallels be-

tween Rome and the ancient Near East so striking that he claims that Roman manumission laws were influenced by those of Mesopotamia, a view that Ernst Levy rejects as “unwarranted in every respect,” adding caustically that “a device which other nations were able to introduce was certainly not beyond the reach of Romans.”**

Bernard Siegel makes an extreme claim for the Third Dynasty of Ur. “There is,” he states, “considerable documentary evidence for manumission and freedom, which when once established, completely freed the slave from the stigma of his former status.”” If this is indeed the case, then the Third Dynasty of Ur ranks as an unparalleled instance of tolerance toward the exslave.

We should also include in this third group the freedmen of India during the Buddhist period. Dev Raj Chanana favorably contrasts their fate with

that of freedmen in Greece and Rome. He errs in lumping Greece and Rome together and paints an exaggeratedly bleak picture of the Roman freedman; and the evidence he cites for the extremely favorable status of the Indian freedman is none too persuasive. He concludes by telling us that the

social integration of the Indian freedman “once he had been manumitted was immediate and complete.”””® I take this with the same skepticism as I do Siegel’s assessment of freedmen in the Third Dynasty of Ur. All these societies, then—Rome, ancient China, India during the period

of Buddha, ancient Egypt, ancient Mesopotamia, and medieval Korea— constitute a special group in which the political and social status of the freedman was relatively most favorable: wala relations were formalized and strong, but not excessively demanding. Citizenship, though perhaps secondclass, was achieved by the freedman, and the stigma of slavery disappeared

within two or three generations. |

(4) In the fourth group of societies, we find that although economic ties of dependence, as always, continued to be strong, there was no institutionalization of the relationship between ex-master and ex-slave. The freedman was, at least in theory, free to go where he pleased. Although agriculture continued to be the economic base, these societies had a strong urban and mercantile character. A considerable proportion of slaves, in some cases the majority, were located in the urban or industrial sectors and performed the full range of occupations. The economic condition of the freedman was determined by the kind of relationship he or she worked out with the former master and by his sectoral location: that is, whether rural, small farming, latifundia or plantation, mining, or urban commercial. There was, however,

254 Slavery as an Institutional Process an interesting and distinctive relationship between the political-legal and prestige status of the freedman on the one hand, and the rate of manumission and condition of slaves on the other hand. The overall rate of manumission tended to be high in this group of societies (freedmen and their descendants made up between 25 and 50 percent of the total population), but the freedman was rigidly excluded from full or even partial citizenship even though he participated fully in the economic life of the community. There

was a strong and persisting stigma attached to freedman status, and this continued for generations. Indeed, the freedman group came to form a sepa-

rate caste. Many were descendants of members of the master class and cherished their ethnic, racial, or class ties, using them as a means of separating themselves from the slave group. Unlike the Islamic group, there was no physical absorption of the children of concubines and no formal recognition of the concubinal relationship. I include in this group the slave societies of ancient Greece, especially

Athens and Delphi; all the slave societies of Latin America except Cuba during the nineteenth century; the Dutch commercial colony of Curacao; and South Africa during the eighteenth century. Although it may seem extraordinary to include Greece here, this is only because of the conventional view (which I| have already criticized) that racial differences between masters and slaves in Latin America differentiate these societies sharply from those of the Greco-Roman world. A significant common feature of this category of slave societies is that despite the large freedman population in the urban areas, there was relatively little economic conflict between freedmen and the free artisan group. This, as we shall see, contrasts strikingly with some of the other groups of societies. There are several reasons for this relative absence of economic hostility. One is the fact that all these societies had continuously expanding urban economies. The demand for skilled and semiskilled labor always outpaced the available supply, so that slave labor did not unduly depress the wages of the free. Second, most of the urban slaves were in fact owned in relatively small holdings by people who were barely coping, or by the free artisans themselves who either employed them in their own workshops or allowed them to hire themselves out. The artisan class, then, did not see the slave population as competitors, but in many cases as a vital part of its labor force. At the same time, in order to motivate its slave working force, it resorted to the technique of manumission with, of course, strong economic ties of dependency. The kind of work the slaves performed also allowed them to

acquire a peculium large enough to buy their freedom. Since the artisan class, then, partially created the freedman class to serve its own interests, it was hardly in a position to resent it. Third, and perhaps most important, was

the attitude toward labor. In ancient Greece, as in preabolition Latin America and eighteenth-century South Africa, all forms of labor were

The Status of Freed Persons = 255 viewed with some contempt: they were banausic. What the free Greek artisan, like his Latin American counterpart, desired more than anything else was to accumulate enough wealth so that he could retire early from the despised artisan crafts. Unlike regions such as the United States South, there was no categorization of some crafts as worthy to be reserved for freemen and others as unworthy and fit only for slaves. In Greece all skilled work, including even architecture, was viewed with contempt.*’ The distaste for labor, including the “banausic” skills, may have been less extreme in Latin America and South Africa, but it was nonetheless present.** It is no accident that the best pieces of eighteenth-century South African architecture that have survived are all the work of freedmen or slaves—or that in both Lima and Buenos Aires freedmen were able to become masters of craft guilds.* The inclusion of eighteenth-century South Africa in this group of societies may seem surprising. Although the fact is not well known, eighteenthcentury South Africa was a large-scale slave system, “a complete microcosm of that of the Americas,” as Lewis J. Greenstein has observed.*? South African slave society did not closely approximate any single New World slave society. Economically, it resembled seventeenth-century Peru and Mexico in

its combination of large hacienda-type farms where most of the slaves worked and in having an extremely important urban center, Cape Town, the

seat of the highly autocratic Dutch East India Company.*! Demographically, South Africa was more like northeastern Brazil with the slave and “free black” population together outnumbering the whites, although the latter constituted a substantial minority of well over 40 percent of the total population.” In the high rate of natural decrease of the slave population it closely resembled both Brazil and the British Caribbean during the eighteenth century.” In its hostility to manumission it closely resembled the U.S.

South and the British Caribbean, sharing with these societies the lowest manumission rates in the history of large-scale slavery.“* Finally, in cultural terms, it was closest to the U.S. South, having a relatively large, settled white population with a strong puritanical tradition; an even sex ratio, allowing for stable family life among the whites; and an incapacity to resist the “dreadful sin” of miscegenation.* However, the small freedman class that did emerge lived a far less con-

fined and oppressive existence than its U.S. counterpart. (South Africa’s barbaric racial policy is very much a product of the nineteenth and twentieth centuries.) While its system of slavery was harsh, in the opportunities and status it allowed the freed blacks it compared favorably with the most open of the Latin American slave societies.*° What is more, the status of the freedmen in South Africa not only closely resembled that of Latin America but did so for much the same reasons. The freed blacks were not needed as a

buffer the way they were in the British and French Caribbean and in Surinam, so this was not the reason for their relatively better status.

256 Slavery as an Institutional Process The condition of the freed slaves in South Africa bears an unusual resemblance to that of their counterparts in Lima and Buenos Aires. First, they were all overwhelmingly urban in location or origin.*’ At the same time, competition from white artisans, though it existed, was never as severe and oppressive as in the U.S. South.** The demographic profiles also are al-

most identical. There were more females; the age distribution showed the same high dependency ratio; there was a disproportionate number of persons of mixed racial origins; the vast majority of manumissions were purchased; and there was a strong correlation between the mode of acquisition and the form of manumission”’ (see Table 10.1 in the next chapter). Finally, the economic, political, and social constraints were about the same in all these areas. While looser than in the U.S. South, they should not be idealized. A small minority of freedmen were able to become modestly prosperous, whereas the great majority lived at or below contemporary standards of poverty. They usually had little or no say in municipal affairs and of course were excluded from important decision-making processes and offices.” Finally, they suffered discrimination resulting from both racial prejudice and their former slave status. There were also laws that regulated their behavior somewhat more severely than that of freeborn persons (although these were never as onerous as they were in the United States). Strikingly similar were the sumptuary laws aimed mainly at the freedwomen: in Lima, they were forbidden “silk, pearls, gold slippers ornamented with silver bells, canopied beds, and rugs or cushions to sit on at church”;”’ in South Africa, jealous white women decided that freed women, by their dress and manner, had become “unseemly and vexing to the public” and in 1765 they were forbidden to wear “colored silk clothing, hoopskirts, fine laces, adorned bonnets, curled hair or earrings.”°* One can understand the vexation over silks, but forbidding a mulatto to walk in public with her hair in curls a hundred fifty years before the invention of the straightening iron was an early and ominous indication of the white South African talent for fine-tuned racial sadism. The South African case highlights two important aspects of the release from slavery. One is the fact that there is not necessarily any association between the rate of manumission and the status of the manumitted. The rate in Buenos Aires was two and a half to eight times greater than in Cape Town, yet the status of freedmen was similar and quite unlike their status in the U.S. South, which had a manumission rate more like that of the Cape. The second important implication of this comparison of freedman status at the Cape and elsewhere is that the status of freedmen, and to some extent the manumission rate, was more sensitive to what Hoetink calls “different sets of

secondary economic, demographic, and social conditions” than to broad macrosocioeconomic configurations.” (5) Let us now move to the fifth group of societies, in which the master

The Status of Freed Persons 257 class was not only racially distinct but constituted a small minority—between 10 and 15 percent of the total population. The overwhelming majority of the population were slaves. There was a strong prohibition on manumitting slaves, masters wishing to do so sometimes requiring legislative permission. Even so, a freedman group did emerge, largely through concubinage. A minority of freedmen purchased their freedom, but this was extremely difficult given the capitalistic nature of these societies, the high replacement value of slaves, and the hostility of planters to the idea. There was no institutionalization of the wala relationship: the large number of slaves and the capitalistic mode of production meant that personal relationships rarely developed between field slaves and members of the master class. The freedman population was demographically peculiar, especially during the early and middle periods of these slave formations. There were far

more women than men and a disproportionate number of children. The male freedmen tended to be older persons, since they were able to purchase

their freedom only after a lifetime of saving, or they had been unscrupulously granted their freedom when they had become too old to be worth their keep. What was most distinctive about this group of societies was that in spite of the hostility toward manumission, the master class viewed the freedman group with some ambivalence. This was partly because of the strong sexual

and illegitimate kinship ties between the two groups. But it resulted primarily from the racial insecurity of the minority master class (see Table 9.1) and its fear of slave revolts. The racially mixed “freed coloreds” were seen as a vital buffer between the masters and ihe mass of black slaves. The freedman class strongly identified with the master class and exploited its buffer status to good effect. Notoriously, freedmen were among the cruelest masters. Because almost all whites were slaveholders involved with the planta-

tion economy, there was not a large group of free white artisans who resented the freedmen as economic competitors. Where such a group existed, as in Barbados and the French Antilles, it was too small to constitute a major source of repression for the freed group, and its interests always took second place to the grudgingly recognized value of the freedman as a racial buffer. With this leverage the free coloreds in all these societies increasingly im-

proved their civil status until they attained full citizenship and had equal legal status with all other free persons. Yet they continued to suffer the stigma of slave and partial black ancestry. All the French and British Carib-

bean slave societies, as well as Surinam (but not Curacao), fell into this group, as did the Mascarene Islands (Mauritius, Reunion, Rodrigues, Seychelles, and their dependencies) during the eighteenth century and the Dutch slave system of the Banda Islands south of Ceram.™*

(6) The sixth and last group of societies is a group with only one

TS )TAT 7 |e we ~Ssil ‘C A2[alway S eS) 3 lon | SsSilgligs S wieal Sac| 2| lw +. < 55 akSiSa | |AS OS

|S/S eee it © 2 | my 5 2:N 7 :| mom |e SO A wae :

.S; rag

> =

Fis DANK E ES ele Foy agg5z ’ |a1|= || :: |SPSS 8 ga! Fa $ “Te 1g FE

:ie iz Da |2+ee|e

oO eee 2a. o i &[~ ae laws ae:: | ae a2iooon |Sis RO)5% eS : sid |gle 22-22 2} | 00 -2er : |~ | ° Plats b~an S52)~~~ a©v: E O09] 8 é

2. Qi | a : os 2jes siSB EER)|2153 igre seit33

one is 7 Ea a | ace , : ~~ , x sf | 33 a); 5 | 2ene

7=Ss|7NS 1112/6 en LARD aSSES lL| |KR KR aA ~

| So lg-1|1 ie2Sq| ::

pom i| So)SSSIS

‘S| ~iloeal il i | Sf iii) isis = SN G5sss55 8 fame |

The Status of Freed Persons 259 member: the slave states of the United States. Berlin tells us: “Once free, blacks generally remained at the bottom of the social ladder, despised by whites, burdened with increasingly oppressive racial proscriptions, and subjected to verbal and physical abuse. Free negroes stood outside the direct governance of a master, but in the eyes of many whites their place in society had not been significantly altered. They were slaves without masters.”°° There was literally no place for the freedman in this slave formation. The South was unusual, too, in the wholly irrational fear of the freedmen who were seen as marginal and “dangerous” persons, an “anomalous caste.” Although obviously harmless politically, “in the white mind the free negro was considerably more dangerous than the slave” and whites “uniformly identified them with the most rebellious.”°° Like witches and other marginals among primitives, the freedmen became classic scapegoats, easy prey for every troubled white free person. Bills to expel all free blacks were repeatedly introduced all over the South, although not all states enacted them. Seven states required freedmen to leave the state, and thirteen made their immigration illegal.’ How do we account for this remarkable situation? One of the causes was the peculiar economic condition prevailing in the South during the nineteenth century, when the freedman group attained some numerical significance: a booming agrarian slave system with an insatiable demand for slaves but no external source of supply, existing within the context of a wider continental economy in which wage labor was the norm. The economic structure, however, better explains the hostility to manumission and its low rate than the hostility to the small freedman class. We have already seen, in our dis-

cussion of the Caribbean cases, that hostility toward manumission did not necessarily imply complete hostility to freedmen. Nor, in this instance, can the racial differences between freedmen and white have been the determin-

ing or even the most important factor. |

The main reasons for the peculiarly oppressed status of the freed class in the South were the unusual demographic structure of the area; the economic fears of the large free white artisan and working class; the absence of a for-

malized wala relationship; the puritanical tradition; and the familial and sexual values of the whites, especially as reflected in their attitudes toward women.

The demographic mix of the South was unusual. The slave population, although large, was always a minority (rarely above a third of the total) and was unusual in reproducing itself. There was relatively less political fear of black slaves than, say, in the Caribbean with its large slave populations and its traditions of massive slave revolts. Thus there was little need for a racial buffer, and the freedmen could not exploit that role. At the same time, the large white lower-class population, especially the growing immigrant group during the nineteenth century, saw the freedmen who converged dispropor-

260 Slavery as an Institutional Process tionately in the urban areas as economic competitors. Unlike ancient Greece and many parts of Latin America in the eighteenth and nineteenth centuries,

the urban economy was not growing so fast that the demand for skilled labor ran ahead of the supply. Nor was there any cultural disdain for skilled labor. On the contrary, the white immigrants came from societies in which there was a proud tradition of skilled work. They were strongly opposed to any lowering of the status of crafts by their association with slaves or freed blacks. Instead, the most low-paying semiskilled activities were soon identified as “nigger work,” while the better-paying skilled crafts were exclusively confined to white workers. In this economic struggle racism became an easy weapon in the hands of the white artisans. But the hostility of the white working class cannot sufficiently explain the unusual status of the southern freedman for, as Berlin has shown, despite their disproportionate location in the urban areas, the majority of the freedmen actually remained in the rural parts of the South. The absence of a formal wala relationship is another factor explaining the freedman’s oppressed status. While ties of economic dependence continued in the rural areas, such ties were difficult to establish in a system where most slaves were field hands in what were highly capitalistic rural firms. There were several more reasons why masters did not wish formally and openly to continue their relationship with ex-slaves even when the latter had been house slaves or concubines whom they had known well. These reasons must be considered independently of their contribution to the absence of a formal wala relationship. One is that southern masters genuinely felt that the presence of freedmen set a bad example to the slaves. In most other slaveholding societies, of course, masters welcomed this example as a means of motivating their slave population to work harder. However, the southern masters chose a different incentive scheme, that of rewarding the slave materially within the context of slavery. Once they had selected this kind of reward system, slaveholders felt obliged to remove freedmen or to depress their status so as to avoid making manumission a competing incentive mechanism. (This is in part why many states demanded that the freedmen remove themselves altogether from the state.) There were other reasons for the demand that the freedmen go elsewhere: the marital traditions of the planter class, their attitudes toward the women of their class, and their fundamentalist religious values. In all other Slaveholding societies the sexual exploitation of slave women was either fully sanctioned by social and religious law (as in Islamic lands and other societies that practiced polygamy and formal concubinage) or the moral system, along with social practice, accommodated such exploitation (as in the Catholic slaveholding societies of medieval Europe and Latin America). In the Protestant slaveholding societies of the Caribbean the church had al-

| The Status of Freed Persons 261 most no impact, this being a part of the general erosion of moral values among the whites, and the high male-female sex ratio among the whites made concubinage not just a biological necessity but in most areas required

practice on the plantations. ,

The U.S. South shared with other slaveholding societies the exploitation of slave women and the inclination of masters to manumit their concubines and children. The intense shame that the master class felt about this sexual relationship was absolutely unique to the South, however. The guilt, with its disastrous consequence for the freedmen, had three sources. First, there was the puritanical tradition, which condemned fornication with the threat of fire and brimstone. Second, there was a highly developed sense of racial purity frequently codified in laws against miscegenation. And third, there was a strong moral commitment to a patriarchal family life, in which the women of the master class were placed on a pedestal and became symbolic not only of all that was virtuous, but as W. J. Cash has argued, of “the very notion of the South itself.” The cult of southern womanhood was of course directly derived from slavery and the sense of racial superiority. Any assault on the dignity and honor of the idolized woman was an assault on the entire system.” But southern males were no less pleasure-loving than the men of any other slaveholding society. Their hedonism, however, conflicted with their religious values, making the southern master alive to a deep sense of sin and wickedness: “‘the Southerner’s frolic humor, his continual violation of his strict precepts in action, might serve constantly to exacerbate the sense of sin in him, to keep his zest for absolution always at white heat, to make him humbly amenable to the public proposals of his preachers, acquiescent in their demands for the incessant extension of their rule.”*’ Equally, his hedonistic exploitation of the slave women was an assault on the integrity of the idolized women, all of whom were constantly reminding him of his wickedness when they were not displacing their bitterness in acts of cruelty toward comely female slaves. The result of all this was that the freed group, with its disproportionate number of mixed-blood members, was a living reproof, a caste of shame, confronting the white males with the fact that they repeatedly violated not only their puritanical precepts but the honor of their women. It was not guilt about slavery that accounts for the exceptional hostility toward freedmen, as Berlin and others claim, or any real fear of them as a political threat, but guilt about their own violation of their own social order. The “zest for abso-

lution always at white heat’ made it imperative that the freedmen be scourged from their midst—or, if not scourged, punished, victimized, and defiled like scapegoats.

Nothing like this had previously existed in the long annals of human slavery.

| ee]

of :ee| Patterns

_ Manumission |

IN DISCUSSING the frequency of manumission, we need to

recall the distinction between intragenerational and intergenerational mobility. The vast majority of societies tended to manumit their slaves over time, replacing them with fresh recruits, so that there often was an intergenerational tendency toward release from slavery when there was little intragenerational manumission. Intragenerational mobility was clearly far more important to a slave population than intergenerational mobility. Over three or four generations the descendants of most slaves were released—but to most slaves the pragmatism of Lord Keynes would have been far more meaningful, namely, that in the long run, we are all dead. The fact that his great-grandchildren might be free perhaps was of some consolation to the first-generation or second-generation slave, but there are limits to the capacity of human beings to postpone gratification. Most slaves in postprimitive societies would have preferred release within their lifetime, and their extraordinary efforts to secure freedom, not just for their children but for themselves, is an impressive testament to the human drive toward independence. Indeed, the most comprehensive body of evidence on manumission in any premodern society, that on Delphi, makes it abundantly clear that, given a choice, human beings will take freedom for themselves at the expense of their children. Keith Hopkins claims that “parents even left chil262

Patterns of Manumission 263 dren behind in slavery to win freedom for themselves as adults.”’ Studies on eighteenth-century slavery in the Americas and in South Africa reveal that

not only did many parents face this tragic plight but many adult children had to leave their parents in slavery.’ There are two kinds of issues involved in the frequency of manumission.

One concerns variations in the rate of manumission across societies. The second addresses the varying rates among different groups of individuals within societies, regardless of the overall rate of manumission. We begin with the latter.

The Incidence of Manumission What factors determine which slaves are released from slavery and which are not? We have already come across most of them in our discussion of the treatment of slaves. The most important variables are sex, status of parent,

age, skill, means of acquisition, color, and residence (where relevant). Tables 10.1 and 10.2 present raw data on a select number of societies. In nearly all slaveholding societies female slaves were manumitted at a higher rate than males, whatever the overall manumission rate, primarily because of their frequent sexual relations with the master or with other free males. Prostitution was also an important source of income for slaves, and in many cases for owners, and a path to manumission; it was usually closed to male slaves, although there were exceptions, perhaps the most notable being the Barbary states from the sixteenth to the eighteenth century, where sodomy was widespread and male prostitution, as well as male concubinage (sometimes involving female owners), was common.° Data on the extent of female slave prostitution are hard to come by, but the literary evidence suggests that while servile prostitution frequently occurred, it was more pronounced in certain areas. It was very common in ancient Greece and Rome and must have accounted for a good part of the peculium used by women to buy their freedom at Delphi; it was also an important way to accumulate the redemption fee in sixteenth- and seventeenth-century Valencia.* Perhaps

the most extraordinary case is that of Cape Town in eighteenth-century South Africa, where the slave lodgings of the puritanical Dutch East India Company became notorious as the best and biggest whorehouse in town.” Another factor favoring the female slave in the manumission process was the mother-child bond, which under slavery was not only stronger than the

father-child relation, but may often have been the only parental bond. Hence mothers tended to be bought out of bondage by previously freed children to a much greater degree than were fathers. Women in their maternal role also had far more numerous opportunities to establish close personal bonds with their owners, whom they might have reared. Probably the most important reason why women were manumitted at a greater rate than

264 Slavery as an Institutional Process Table 10.1 Percentage distribution of those manumitted, by color, gender, age, and form of manumission in selected cities.

Buenos Aires Bahia Paraty Lima Mexico City 1776-1810 1684-1745 1813-1853 1789-1822 1580-1650 1580-1650 Color

Negro 51.3 54.4 80. | 50.6 — ~~ Mulatto 48.7 45.6 19.9 49.4 — Total (1,316) (945) (657) (320) — — —

Gender

Female 58.833.1 66.932.7 67.334.5 65.532.3 67.7 38.5 61.5 Male 41.2

Total (1,482) (1,150) (686) (325) (294) (104)

0-5 14.6 9.2 — 22.0 36.0 41.5 6-13 71 35.6 — 19.8 15.9 14-45 67.0 52.3 — 43.3 35.512.3 33.9

Age

46 and over 11.3 2.9 — 14.9 12.6 12.3 Total (937) (763) — (268) (214) (65) Form of manumission

Gratis 29.3 —— 31.5 Purchased 59.8 46.026.1 31.433.8 47.8 39.3 36.4

Conditional 10.9 22.5 42.5(299) 18.4 (107) 24.3 Total (1,356) —— (561) (325) souRCE: Lyman L. Johnson, “Manumission in Colonial Buenos Aires, 1776-1810,” Hispanic American Historical Review 59 (1979): 262. Copyright © 1979 Duke University Press, Durham, N.C.

men is that in all societies so-called free women were far more dependent than free men. Masters took fewer nsks in losing the services of female Slaves by freeing them than they took in freeing male slaves. Despite all this, in many cases, especially in ancient Greece and Rome, the Americas, and South Africa, women were obliged to pay full replacement costs for their freedom.°

The status of parents has already been discussed at length,’ and we saw that there was considerable variation in the ways in which mixed (free/slave) parenthood influenced manumission. Apart from the Islamic lands, only a small minority of the children of masters by their slaves in the advanced slave systems received their freedom. This remained true even where, as in many parts of Latin America and South Africa, such children constituted a disproportionate number of those who were actually manumitted. In most slaveholding societies slaves who managed to acquire some skill, or who were already skilled when enslaved, were better able to accumulate

~-_ a aQsiy F SELF aAnaanevoygs

sToerse NAMM TOM oa AS £8 v o ”A EE QQ, Saat woe” ©) -_—pen Sz. EE : Ss Ssirranaetrnoa = 2S 8 9 — ot a NAO et o SSE 35ES _ "Sb SES 2 os § 5 3 =|

x £3 Ss 4 “ese” . Oo 3 Sr S oe) 2. OTF —_ ma 2s srs js ma =aiAee Acs TaetT CO r?2 ~ ~~ ww ™~ ja n FE & . nw Wk as Sy lan NY -f} x Cond Ow HY —N =~ a6& op |S — 8+ .an Fay == “nF 28 es /omoamons + Ooa 8 £§ |=oanarale ~ £8 rm fo cS 30

waeke flfaeBVDsse ev sonnn=—-ono = VRP IMNaAMNOAMS oa -«ot ¥

_>3

S SG lrnawnonmonra | ¥ ~=

& aml , > =v >OY& Fla "A ge & a3 —_ Foote AaNH o a3° ~ ° 8 >

384 BS) = ‘rc

= toss

Ds SalYo | manasa = ..S73 m= oU n Le —~ |jo %

= = S385 5 = ~ o XS > & AS jaunt anoaer | PCa s pn ec £1 \o — & oe) =S Vik,Sy mH sore SBS > 3°0 — 1°) GS Tan ntnnrnr o Fr>ss B = nm Ss zeae 7) SéaZo v 3 wy re 2 SSel( tans0| an A apo ™ ~ |UO Oo % Oo EL Mm COMO DH +t nm |fg BOF e

y BS |peweonnrnrwoan S°G6 eB

ee SAR lLanwowennn HH |82%gm Gq

q ~ZSec > yoes2 ‘2

em AAs438 — OS24 -s aCebw a ot, 8S Fy oAy oo |r S325 7 33+¢ o € a6 . -™ a5 oO:3 = SG’WVvRIMNMAMMOMAO ~ | 55032 &

2 )|e6a Ee &

w .228g os oe Seciagesd see veo Ss) Sek

SP)sSimrot sslaaz 5PA9 = Wm CO moOe22 8 ee ee oeays oe on oeooGes S esloRnsnaa om SLA = aNd puns: ment, O72, » an

freedom, 94; in U.S. South, 94-97; and debt-bondage, 404n132 valry 95. 97: of Hebrews (Palestine): intrusiveness of slav- chivalry, 95, 97; of Master, 96; and rec-

ery, 40-41. enslavement practices, 41. ognition, 99-100: caliphs and, 308-311 125, 275, 404n 132; killing of slaves, 192; Honor payment. 82

manumission rate. 271, 275 Hopkins, Keith, 265; on Augustan Lares

Hedeby (Scandinavia), 154 cult, 69; on slavery and wealth in

Hedonism. 261 Rome, 92; on Delphic manumissions, Hegel, Georg: on contradictions to total 230-23 L, 238, 262-263, on race and . power, 2; on political-psychological as- ManNuMssion, 277-278; on eunuchs in

pects of slavery, 11: on dialectics of ie agra ote nha 315, 317,

slavery, 97-101, 392n115; on limits, » 319, 320, 322, 327, on migrations 299: on dependence of dominator. 336 of Itahan poor, 375n64

Index 495

Horace, 89, 252 Igala kingdom (West Africa), 195, 316 Household slaves, 427n62; selling, 175; Igbo-speaking groups (Niger), 233

manumission, 232, 270-271 Ila (Central Africa): rituals of enslave-

Hovamol (Icelander), 82 ment, 53; marks of servitude, 60: Hsuiung-nu (China), 108 status inheritance pattern, 147; change Hubris (Greek concept), 389n58 of master, 203-204; freedman status,

Hughes, G. B. J., 22 251

Human parasitism, 81, 334-342, 453n1; Iliad (Homer), 121, 230 and domain of freedom, 14, 340-341; as Hlyrians, 116 relation of domination, 173, 334-339. Imbangala (Angola): rituals of enslavevariables, 206; personal satisfaction in, ment, 53; fictive kinship system, 63; sta336; mutualism, 336-337, 338; institu- tus inheritance pattern, 136; concubinal tional aspects, 337, 341; gift exchange manumission, 232; manumission rate,

theory, 341; on systemic level, 341; 280 Rapoport’s theory, 336, 453n3 Imerina (Madagascar), 280 Human sacrifice, 428n84; alternative to Immigration/emigration rates, 132,

enslavement, 107; in postmortem manu- 414n30

mission, 220-228, 433n45; principle of Imperial China, 24, 43 |

gift exchange, 221, 224, 227; substitu- Imuhag (Tuareg group): fictive kinship tion of animals, 221, 223; at death of system, 64; change of master, 203 master, 224, 226; as capital punishment, —Inalcik, Halil, 308-309

429n90. See also Potlatch ceremony Incentive, see Reward

Hundley, Daniel R., 95-96 Incentive payment, 283-284 Hunter-gatherer societies, 18 Incest, 329

“Hymn to Rome” (Melinno), 90 Indentured servants, 9, 25, 183 India: caste in, 49; marks of servitude, 60:

Iban (Borneo): status inheritance pattern, nature of honor, 80, 86; status inheri-

147, manumission rate, 271 tance pattern, 140; Mauryan period, Iberian Peninsula: enslavement of war 140; slaves as dowry, 167; racial prefer-

captives, 114; slave trade, 160, ence in slaves, 178, 311; treatment of

Ibn Fenari, 313 slaves, 207; manumission ritual, 216:

Ibn Khaldun, 310 postmortem manumission, 221; concuIbn Zurdadhbih, 123 binal manumission, 230, freedman staIbos (Nigeria): marks of servitude, 58; tus, 253; manumission rate, 281; eunature of honor, 83; enslavement prac- nuchs, 320, 422n20 tices, 120, 126, 128, 423n26; status in- Indian Ocean area: slave trade, 150, 151;

heritance pattern, 137; slaves as bride- plantations, 181 payment, 166; concubinal manumis- Indians, see Northwest coast Indians sion, 230; manumission of Osu, 238: Indonesia, 469n75. See also Banda Is-

manumission rate, 271 lands; Nias Ibrahim Pasha, 313 Ingram, J. K., 21

Iceland: rituals of enslavement, 55: nature Initiation ceremonies, 53-54, 191 of honor, 79, 81; ransom and warfare, Innate slavishness, 68, 420n12 107; status inheritance pattern, 141: Institutional process, 13; enslavement of slave trade, 152, 157: bride-payment, “free” persons, 105-131; enslavement 167; slaves as money, 168, 169: racial by birth, 132-147; acquisition of slaves,

ideal, 177; crimes against slaves. 194: 148-171 |

crimes of slaves, 196-197, 200; self- Interbreeding, 134, 135, 137

defense, 200: postmortem manumission, — Intermarriage, 137-138. slave and free,

220; economic role of slaves, 386n 18: 64, 409n25; and status inheritance pat-

slave population, 459nI1 tern, 137-138, 142, 409n25: slave assim-

Idonei (skilled slaves), 270 lation by, 279

496 Index Interregional movement, of slaves, 165, manumission rate, 272, 276-277, 278,

166 282, 288; military slaves, 288, 448n31,

Intisap (tacit relationship), 309 467n63; elite slaves, 299, 308-314, 315,

Iranians, 310 316, 321, 326, 449n32, 451n64; concubiIraq: treatment of slaves, 174; tenant nage, 402n107; homosexual slaves, slaves, 181; slave unions, 187; killing of 421n18; sexual relations with slaves, slaves, 190; freedman status, 248; man- 440n3; manpower needs, 450n37. See

umission rate, 274; palace slaves, 332; also Arabs

slave population, 467n63. See also Israelites, 75

Mesopotamia Istanbul, 266

Ireland: Viking raids, 116; status inheri- Italian colonies: enslavement of war captance pattern, 141; slave trade, 152, 154, tives, 114; slave trade, 152; slave plan-

156; slaves as money, 168-169 tations, 181 |

Iroquois (North America), 432n15 Italy: penal slavery, 128; killing of slaves, Islam: rationalization of slavery, 72; social 192; freedman status, 243; manumission

inferiority of slaves, 93; equality of rate, 274, 296. See also Italian colonies: slaves under God, 93; sense of honor, Roman Empire; Romans 97; enslavement of Christians, 117, 196; — Izard, Michel, 38

enslavement of fellow Muslims, 118, Izgoi (aliens), 43 276; and slave trade, 150, 154, 157; tes-

tamentary manumission, 222; post- Jahore (Malay), 200 mortem manumission, 227; and manu- Jakhanke (Fulani subgroup), 461n26

mission rate, 276-277 Jamaica: postemancipation, 2-3, 5-6; Islamic law: intrusive conception of social slave naming rituals, 57; kin terms, 65; death, 41; enslavement of coreligionists, religion and slavery, 72; enslavement

41, 118; enslavement for debt, 125, practices, 113, 120; reproductive and 404n 130; penal slavery, 128; concubi- mortality rates, 133, 407n3, 477; slave nage, 145; status inheritance, 145; mar- population, 133, 134, 470n77; status in-

riage, 187, 188; change of master, heritance pattern, 146; internal slave

204-205 trade, 165-166; absentee slave owners, Islamic societies, economic dependence, 180; slave unions, 187; concubinal man11; marks of servitude, 58; honor, umussion, 231: incidence of manumis92-94; ransom and slaves, 107: enslave- sion, 269; manumission rate, 273, 291. ment of war captives, | 14—115; internal parasitism, 338; age at manumission, kidnapping, 117-118; sexual bias in en- 441n11; military slaves, 446n91. See slavement, 121; tribute in slaves, 123; also Caribbean; Worthy Park Planta-

debt-slavery, 124-125; penal slavery, tion 128; social reproduction, 133; status Janissaries, 449n34; natal alienation, 8, inheritance patterns, 138-139, 145; and 311-312; recruitment, 123, 308, 310, slave trade, 150, 157; household slaves, 311-312; race, 248; manumission, 313, 174; racism, 176, 178; skilled slaves, 450n52; honor, 314 180; slave unions, 187, 188; crimes Japan: enslavement by kidnapping, 117;

against slaves, 195-196; crimes of penal slavery, 127; self-enslavement, slaves, 196; treatment of slaves, 199: 130, 131; status inheritance pattern, change of master, 202-203, 204; testa- 141; killing of slaves, 191; invasion of mentary manumission, 222, 239; post- Korea, 289. See also Oriental societies mortem manumission, 227; concubinal Jefferson, Thomas, 95 manumission, 228, 229, 230, 231, 239; Jelgobi (Upper Volta), 83 manumission by adoption, 232; political Jesuits, 427n73 manumission, 235; contractual manu- Jews, 40, 145, 196; in slave trade, 152: mission, 239; freedman status, 241, 250, manumiussion rates, 275-276 252; ethnicity and manumission, 268; Jihad (warfare), 115

Index 497 Johnson, Lyman L., 264, 273; on manu- binal manumission, 228, 232; adoption, mission in Buenos Aires, 266, 268, 290 233. See also Matrilineal societies; Pat-

Johnson, Samuel, 94 - rilineal societies

Jones, J. Walter, 138 Kitab (payment method), 239

Jordan, Winthrop D., 6-7 Klein, Herbert S., 162, 165, 403n113 Judaism: law and judgment ethic, 71; Kojéve, Alexandre, 98, 99 secular and sacred slaves, 72; manumis- | Kongmin Wang (Korean ruler), 143

sion, 275 Kongo (Zaire), self-enslavement, 130;

Judgment, ethic of, 71, 74, 75 slave lineages, 136; manumission

Jurisprudence: Anglo-American, 20, 21; ceremony, 215; manumission rate, rights and duties, 20-21, 22; and law, — 280

23, 36 , Kopytoff, Igor, 33, 248, 296, 438n27

Justification in Paul’s theology, 70 Koran, 41, 203, 235, 239

Justinian (emperor of Rome), 231, 239 Korea: context of slavery, 39, 42; occupa-

Jutland (Denmark), 244 tional castes, 49, 50, 86; prisoners of Juvenal, 91, 178, 421n15 war, 108; enslavement by kidnapping,

117; enslavement for tribute, 123; under

Kachin (Burma): rituals of enslavement, Mongols, 123, 288, 425n44; debt54, 55; manumission by adoption, 233 slavery, 125, 404n133; penal slavery,

Kafur (Muslim leader), 316 | 126-127, 405n146; self-enslavement,

Kahn-Freund, Otto, 29 130; status inheritance pattern, 143; Kallikratidas, 111 slave estates, 181; slave peculium, 183, Kano (Nigeria), 157 184; ownership of land, 183, 184; selfKant, Immanuel, 80 defense, 200; political manumission, Karebe (Africa), 230 235; freedman status, 252, 253, 425n44-. Kasenje (Angola), 53 manumussion rate, 271, 274, 278, 285, Kassi tribe (northwest coast of America), 287, 288-289; reenslavement, 287, 289;

84 | decline of slavery, 289; under Sejong

K’atorshniki (slaves), 59 and Sejo, 289; elite slaves, 299; slave

Kel Ahaggar (Tuareg group), 139 population, 468nn69,70, 469n73. See

Kel Gress (Tuareg group), 4, 139 also Oriental societies

Kentucky, 59 Kory6 period (Korea): context of slavery, Kenya (East Africa): racism and slavery, 39, 42; moral connotations of slavery, 93; runaways and recapture, 94; en- 42; status inheritance pattern, 143; slave

slavement by kidnapping, 117 estates, 181; slave peculium, 183, 184; Kerebe (Tanzania): ransom of slaves, 107; freedman status, 425n44; slave populaenslavement practices, 120; release cere- tion, 468nn69,70

mony, 215 Kruyt, Albert C., 85, 203, 396n46

Keynes, Lord, 262 Kuka market (Chad), 169 Khoikhoi (South Africa), I 11 Kul (elite slave), 308

Kidnapping, enslavement by, 115-122 Kumasi (Ghana), 462n35 Killing of slaves: legal sanctions, 190-193; Kungahalla (Sweden), 156

correlates, 193; by third parties, K’unlun slaves (China), 422n20 193-194; by slaves, 196. See also Kuo Mo-jo, 108

Human sacrifice Kush (North Africa), 123

Kin-based societies: dependence in, 19; Kuwait, 93 naming ceremonies, 55; fictive kinship, Kwakiutl Indians (northwest coast of 62-65; adoption, 63-64; sexual exploita- America): social death, 39-40; killing of tion, 64-65, 232; religious institutions, slaves, 191

66-76; enslavement methods, 113; treat- © Kwanyama (Angola), 53 | ment of slaves, 174; sex ratio of slaves, Kwararafa kingdom (Nigeria), 123 199: manumission ritual, 215; concu- Ky, Dang Trinh, 86

498 Index Laberius (Roman writer), 77 Lévi-Bruhl, Henri, 7, 38, 178 Labor, free: in slave societies, 33-34; sale Levine, Lawrence W., 74

of, 366n21 Lévi-Strauss, Claude, 18, 19, 323, 325

Lactation practices, 133, 162 Levitas (unworthiness), 90-91

Lamson, H. D., 370n26 Levy, Ernst, 253

Land: in peculium, 182, 183; in lineage- Lewis, Monk, 65 based societies, 241; and wealth, 371n41 — Liang (“‘good” status), 86

Lang, Olga, 370n26 Liao (Szechuan), 419n9 Langobards (Germanic group), 217 Liberation beer, 218

Lares cult, 68, 69, 215-216 Libya, 157 Large-scale slave systems, 353-364. See Liebeschuetz, J. H. W., 216 also Slave populations Lima (Peru), 245, 256, 289

Lasker, Bruno, 399n75 Liminal incorporation, 45-51, 293

La Tene culture (Celtic group), 149 Lind, Joan Dyste, 145

Latifundia slaves: free farmers and, 34, Lindkvist, Thomas, 243 372n49, degradation of, 64; cults, Lineage-based societies, see Kin-based so68-69; demand for, 116; manumission, cieties; Matrilineal societies, Patrilineal

282, 284 societies

Latin America: naming rituals, 56-57; Little, Kenneth, 181 branding of runaways, 59; sense of Ljufvina (Russian princess), 177 honor, 97; change of master, 202: con- Lolos (China): marks of servitude, 58; cubinal manumission, 231; conditional racism, 176, 420n10; treatment of

manumission, 245; freedman status, slaves, 178; freedman status, 250; 254, 255: incidence of manumission, manumussion rate, 271 266, 267, 269; manumission rates, 272, Lombards (Italy): status inheritance pat284; nonplantation areas, 427n73. min- tern, 141, 144; granting of sanctuary, ing areas, 442n35. See also names of in- 202

dividual countries Louisiana, 446n94, 483

Lauffer, Siegfried, 36 Lovejoy, Paul, 41, 438n27

282 Lowie, Robert, 18

Laurium mines (Athens), 181, 187, 198, Low, Brooke, 147

Law: slave versus indentured servant, 9: Lozi (Zambia): marks of servitude, 58;

conceptions of property, 21, 28-32; manumission rate, 271; slave populaslave as person, 22-23; confusion with | tion, 347 jurisprudence, 23; dominium, 30-32; Luanda (Angola), 466n54 and tradition, 36; criminal as slave, 43. Lundstedt, Anders V., 20 courts and public opinion, 87; marriage, Luvale (Zambia): ransom and warfare,

186-190; murder of slaves, 190-193. 107; manumission rate, 280

crimes against slaves, 193-196; crimes

of slaves, 196-197; treatment of slaves, Maccube (slaves), 83, 84 198-199; slave as active agent, 199-205; | MacDowell, Douglas M., 29 and manumission, 210-211, 222-223, Machiavelli, Niccolo, [8 228; freedman and, 244, wala relation- Mcllwraith, Thomas F., 39 ship, 244, 252. See also Chinese law, Madagascar, 136, 144. See also Merina Hebrew law; Islamic law; Roman law Madeira Islands, 285; source of slaves,

Leach, Edmund, 323 116-117; manumission rate, 274, Leadership, 18 285 Leeward Islands, 275, 478 Mafakur (slavery), 45, 46

Lencman, J. A., 12! , Magrizi (Arab geographer), 124 Lesbos (Greece), 111 Mahabharata (Hindu epic), 320 Lesser Antilles, 60 Mahomet, Sultan, 308 Leviathan (Hobbes), 10 Maimonides, 41

Index 499 Majorca: manumission rate, 274; slave society, 256, 257, 277-278, 294,

population, 459n10 295-296; and economic conditions, 259,

Malar, Lake, 154 279, 281-287, 295, 434n58, 465n52; Malay: injury to slaves, 196; self-defense variables, 262-263, 264, 272-277,

in, 200; granting of sanctuary, 201; 278-279, 295: gender and, 263-264,

slave raids, 400n85 , 279; parenthood and, 263, 264; race Malaysia: internal kidnapping, 117, 118; and, 264, 268, 272, 273, 277-278; age ~ debt-slavery, 124; racism, 176; crimes and, 264, 266-267, 441n11; form of

against slaves, 195-196 manumission and, 264; level of skills

Male concubinage and prostitution, 263 and, 264-266, 270; means of acquisition

Mali, 249, 271 and, 267; ethnicity and, 267-268; rural Malik Ambar, 422 versus urban, 269-271, 274, 282-284. Malikis (Muslim legal school), 188 societal, 271-278, 294-295; from MurMalindi (Kenya), 94 dock’s world sample, 271-272; level of

Malinowski, Bronislaw, 211-212 miscegenation and, 272-273; in modern Mamluk kingdom (Egypt), 308, 310, Slave societies, 273; religion and,

450n51 273-277, 282; in large-scale societies,

Mamluks: status inheritance pattern, 145, 274-275; domestic assimilation and, 312, 314, 331, 450nS1: freedmen, 248, 279; in lineage- and kin-based societies, 310-311; favored status, 299: recruit- 279-280; domestic exclusiveness, ment methods, 310; mawali, 310-311; 280-281; incentive payment and, » honor, 313; homosexual use, 421n18 283-284; in plantation systems,

Mana (power), 51 284-285, 286; military affairs and, Manchu (Manchuria): absenteeism, 180, 287-293, 295 424n34; manumission rate, 271 -Manumissio vindicta, 236, 238, 239,

Manning, Patrick, 162 435n73 | Manumissio censu, 219, 235, 236 Maori (New Zealand): pollution norms, Manumission: sacral, 67-68, 237-238: of 51; enslavement of war captives, 113;

penal slaves, 127; and social conse- manumission rate, 27 1 . quences, 132, 133; meaning, 209-214, Margi (Nigeria): liminal incorporation, 340-341, 431n1; conceptual problem, 45-46; occupational castes, 49; ransom 211; gift-exchange theory, 211-214, 215, and warfare, 107; status inheritance 217, 218, 243, 294; origins, 214, 340; re- | Pattern, 137; killing of slaves, 19) lease ceremonies, 214—219, 237, 450n52: arginality, see Liminal incorporation Marks of servitude, 51-62, 128, 215,

modes of release, 219-239, 294; post- 381n94 yin

mortem, 219-228, 433n45; testamen- Maroons (Jamaica), 446n91

tary, 222-223, 225-226, 227, 237, 239, Marriave. , d ficatihat 435n73; concubinal, 228-232, 239, 251 Be427n56ng oe ateeee

; a y payments, ; ,

° - , , 49: to slaves, 131, 409n25, 411n68: bride

261, 282; by marriage, 228, 252; by and dowr ts. 166-167. 187

adoption, 232-234, 238; political, 188; and treatment of slaves, 189-190; 234-236, 287; by collusive litigation, religious sanction, 189; manumission

236-237; skills of slave and, 242, by, 228, 229, 252: endogamous, 254-255, conditional and uncondi- — 280-281; formariage, 411n68. See also tional, 245-246; selective, 246; somatic Bride-price; Intermarriage

theory, 268-269; incentives, 285, Martinique, 275, 480 444n65; military, 287, 289-293; mass, Marx, Karl: on power and domination, 288, 289, 290, 292; of janissaries, 313; 2-3; on forms of social life, 17, 20; on

by death of master, 433n41 concealment of power in capitalism, 18,

Manumission rates: and demand for 19; on wage labor versus wage slave, Slaves, 132, 281-282; of children, 25, 370n30; on freedom through work 145-146, 267, 269; and size of slave and labor, 98, 99; on equal exchange, population, 255; and freedman status in 713

500 Index Maryland: status inheritance pattern, 138, Medina (Saudi Arabia), 150 409n25; manumission pattern, 292; run- | Mediterranean area: enslavement prac-

aways, 447n97 tices, 114, 115, 116, 117; child exposure,

Mascarene Islands (Indian Ocean): en- 129; slave trade, 150-152, 154, 157-158, slavement of war captives, 114; slave 160, 170-171; slave plantations, 181; plantations, 181; concubinal manumis- freedman status, 243 sion, 231; freedman status, 257; manu- Meillassoux, Claude, 33, 38, 148-149

mission rate, 274, 285 Melior condicio (inheritance of higher

Massacre, 106, 107, 120 status), 144

Master: change of, 201, 202-205; death of, Ménage, V. L., 313

224, 226; defined, 334-335 Mende (Sierre Leone): marks of servi-

Master-slave relationship: and domina- tude, 59; nature of honor, 83; old tion, 2, 336; maintenance, 3-4, 205; in- slaves, 88; status inheritance pattern, timacy, 12; personal relations, 12, 13; 136; bridewealth, 167, slaves as money, and power, |8—19; symbolic control, 37; 168; household slaves, 175, 419n4; ten-

private and public determinants, 172; ant slaves, 181; slave unions, 187; conintrinsic factors, 173-181; human vari- cubinal manumussion, 230; freedman ables, 206; after manumission, 240-247; status, 248; manumiussion rate, 271: in freedman status, 247, 249, 250, 253, slave population, 462nn3 1,34 294: with elite slaves, 307, 308; versus Mendelsohn, Isaac, 109, 234 patron-client relationship, 309-310. See | Mepone (change of master), 203 also Conditions of slavery; Human par- = Merina (Madagascar): status inheritance

asitism pattern, 136, 144, 280; manumission

Materialism, and slavery, 11, 19 rate, 274, 280-281

Matrilineal circumvention, 279-280 Mesopotamia, 299; slaves as foreigners, Matrilineal societies: adoption, 63-64; sta- 40, 42; sources of slaves, 44, 403n 108; tus inheritance patterns, 136, 138, 139, marks of servitude, 62; sense of honor, 185, 464n44; marriage, 187; manumis- 97; prisoners of war, 107, 109-110, sion, 229, 272-273, 279-280; freedmen, 395n26; piracy, 115; sexual bias in en-

241, 251 slavement, 121; debt-slavery, 125,

Mauritania (West Africa), 159 404n 131; sale of slaves, 175; slave

Mauritius (Mascarene Islands): enslave- unions, 187; concubinal manumission, ment of war captives, 114; freedman_ 230; manumission by adoption, 234; status, 257; slave clothing, 381n91; slave temple slavery, 238; freedman status,

population, 467n62 253; defining slavery in, 371n39. See

Maurya period (India), 140 also Babylonia; Iraq; Third Dynasty of Mauss, Marcel, 211; on honor, 78; on ide- Ur

ology of prestation, 212-214; on manu- Metics, 281, 302

mission rituals, 218, 236—237 Métraux, Alfred, 52

Mawéali (freedmen), 310, 311, 436n!1 Mexico: naming rituals, 56-57; social re-

Mayan (slave), 233 production, 133; status inheritance pat-

Mbanza Manteke (Kongo): status inheri- tern, 146; manumission, 245, 275; hacitance pattern, 136; release ceremony, enda farms, 255; economics of slavery,

215: domestic slaves, 464n41 471n86. See also Aztec Indians

Mecca, 150 Middle Ages: abolishment of slavery, 44; Medieval period: social death, 39; pagans penal slavery, 44, 127-128; rise of Protas enemies, 41; rituals of enslavement, estantism, 72; enslavement methods, 52; branding of slaves, 59; kidnapping, 114; church, 189, 201; concubinal man116; slave trade, 152-157; killing of umission, 231; wala relationship, 243; slaves, 192: manumission, 230; freed- freedmen and descendants, 249; manu-

man Status, 243, 250; incidence of mission rates, 282; sources of slaves, manumission, 268; manumission rates, 398n71; marriage of serfs, 411n68. See

275, 278, 282, 284, 285, 288, 296 also Medieval period

| Index 50l Middle East: slave trade, 150; change of Mubika (slave), 136 master, 202; manumission modes, 228 Muhammad (prophet), 227, 232, 316,

Miers, Suzanne, 33, 248, 296, 438n27 451n61

Miles, J. C., 230 316 Military capacity, 33 Muhammad Kurra, 316

Migiurtini Somali (Somalia), 137, 434n50 =Muhammad-el-Fadhl (ruler of Bagirmi),

Military manumission, 287, 289-293 Mui-tsai institution, 129, 370n26 Military slavery: origins, 39-40, 467n63; Mukataba (slave), 239 manumission and, 288-293; racial as- Miikatebe system, 266 pects, 446nn83,94; recruitment of elites, | Mulattoes, 61, 62, 268

448n31 Murad (Turkish leader), 316

Mill, John Stuart, 10 Murder, see Killing of slaves

Miller, J. C., 136 Murdock, George P., 139, 345 Minas Gerais (Brazil), 483; branding of Murdock world sample, 199, 271, 272.

slaves, 59; internal slave trade, 165: 350-352

manumission rate, 289-290; mortality Muscovy, 43-44

rate, 442n35 Muslim East Africa, see East Africa Ming dynasty (China): eunuchs, 315, 324, | Muslim Spain: enslavement of war cap-

325; role of free women, 325, 326 tives, 114; slave trade, 152, 154: tenant Mining areas: manumission rate, 270, slaves, 181; manumission rate, 274 289-290; mortality rate, 442n35 Muslim traders, 154, 159, 168. See also

Mintz, Sidney W., 25, 370n30 Arabs; Islamic societies

Miscegenation, 61, 146, 255, 261; and Mutilations, 59 manumission rate, 261, 272-273 Mutualism, 336-337, 453n1, 455n3 | Mitamura, Taisuke, 315, 324, 325

Mitchell, J. C., 280 Nagaoundere (Cameroon), 274

Mithras cult, 69 Namath, Joe, 24

Moluccas (Spice Islands), 99, 231 Names: servile, 55, 380n75; in kin-based

Mombasa (Kenya), 94 societies, 55; surnames, 56-58; followMoney, slaves as, 167-171 ing manumission, 56; African, 56, Mongols, 123, 288, 425n44 380n80; of saints, 57-58; features de-

Monteil, Charles V., 249 noted, 58; of slave wives, 189

Moors, 145, 152 Naming ceremonies, 53, 54-55 Moral connotations of slavery, 42, 47-48, Narcissus, 301, 307

455n3 Natal alienation: and heritage of ances-

Moral responsibility of slave, 22 tors, 5-6; use of term, 7-8; perpetual Moramachi period (Japan), 117 | nature, 9; flexibility, 32; symbolization, Morocco (North Africa): piracy, 117; 38; of native populations, L11; janisslave trade, 157; manumission rate, 277, saries and, 311-312 443n49; boy-markets, 421n18; slave Natality, restoration of, 235

populations, 460n 15 Ndembu (Angola), 37

Morrow, G. R., 87, 140, 192, 194 Ndizogu (Nigeria), 138 Mortality rates: and enslavement prac- Nduwanga (Mende slave group), 83 tices, 132, 133, 397n55:; in slave trade, Near East: rituals of enslavement, 55: 159, 162, 163; in slave populations, marks of servitude, 58-59; prisoners of 417n64, 465n51; in mining areas, war, 107, 110; piracy, 115; penal slav-

442n35 ery, 126; exposure and sale of infants,

Morton, Roger F., 93 129; status inheritance patterns,

Mossi (West Africa), 60 144-146; slave trade, [50; slaves as Mother-surrogates, 50 dowry, 167; slaves as money, 168; ten-

Moyd, Olin P., 75 ant slavery, 181, 425n41; slave pecu-

Mozambique, 160, 228, 466n55 lium, 184; killing of slaves, 191; concu-

502 Index Near East (continued ) war captives, 113; status inheritance binal manumission, 231; freedman sta- patterns, 136; slave trade, 149, 156; killtus, 241, 253; manumission rate, 284 ing of slaves, 191; potlatch ceremony, Negroes: prejudice against, 176, 179, 250, 214, 222; sacrifice and manumission,

4? 1n16: free, 259. See also Black 221-222; manumission by adoption,

Americans 234; parasitism, 337; escapees/return-

Nepal: slavery and caste, 51; granting of ees, 432n15. See also Kwakiutl Indians;

sanctuary, 201 Nootka

Nero (emperor of Rome), 307, 308, 317, Norway: nature of honor, 82; slave trade,

326 152, 154; acquisition of slaves, 157: self-

Netherlands, 44. See also Dutch defense in, 200; manumission ritual,

New England, 278, 446n94 216, wala relationship, 244; freedman New World, see Americas status, 249; economic role of slaves,

New York, 138 386n18 New Zealand, see Maori Nubians, [23-124

Nias (Indonesia), 39 Nursemaids, 50, 88 Nicomedes of Bithynia, 123 Nwachukwu-Ogedengbe, K., 233 Nieboer, H. J.: on defining slavery, 21, Nyinba (Nepal), 201 369n18; and treatment of slaves, 181.

402ni02 Obedience, ethic of, 74, 76

Nietzsche, Friedrich Wilhelm, 78 Obsequium (claims on freedmen), 242 Nigeria: nature of honor, 82, 97; internal Odysseus, 87, 121]

kidnapping. 117-118; treatment of Odyssey (Homer), 121, 230 Slaves, 175; political eunuchs, 316. See Oikogeneis (house-born slave), 175

also Aboh; Hausa; Ibos; Margi; Tiv Okigwi (Nigeria), 138

Nikaya, Mahima, 207 Olivecrona, Karl, 20

Nkundu (Zaire): ransom and prisoners, Operae (claims on freedmen), 242, 244,

107; slaves through bridewealth, 167: 283 concubinal manumission, 230; manu- Oriental societies: inheritable slave status,

mission rate, 27] 9; sense of honor, 86; prisoners of war, Noburu, Niida, 141 107; penal slavery, 126-127; exposure

Nonslave persons: sale of, 23-26; work of infants, 129; slave estates, 181: treatethic in large-scale slave systems, 33- ment of slaves, 199. See also China:

34 Japan; Korea

Nootka (northwest coast of America): rit- Origo, Iris, 201 uals of enslavement, 55, 380n79; post- Ormerod, Henry, 115, 116

mortem manumission, 222 Ornaments, as marks of servitude, 58

Nordic peoples: slave trade, 154; freed- Osterweis, Rollin G., 94-95

man status, 243-244 Osu (Ibo cult slaves), 238 Norsemen, 140, 156 Otho (emperor of Rome), 307

Norsus (Lolos), 468n67 Ottoman Empire: piracy, 117; enslave-

North Africa: enslavement of war cap- ment for tribute, 123; recruitment of tives, 107, 114; Barbary pirates, 117; slaves, 248, 312; elite slaves, 299; janisslave trade, 157, Christian captives, saries, 308-309, 313, 449n34

402n 107 Ottoman Turks, 178

North America, southeastern, 107 Outcastes, 48-50, 435n81 Northern dynasties (China): prisoners of Outsider status, 7

war, 109; freedman status, 252 Ownership: concepts, 20; and property,

Northup, Solomon, 392n121 21-22, 28, 371n43; absolute, 28, 29-30, Northwest coast Indians: marks of servi- 31-32

tude, 60; hereditary slavery, 84; treat- Oxenstierna, Eric, 169 ment of prisoners, 107; enslavement of Oyo (Nigeria), 123

Index 503 Paekchong (Korean outcastes), 50 Peter the Great, 43 Paganism, 41, 47, 226, 227-228 Petrochitos, Nicholas, 90 Palaestra (character in The Rope), 12 Petronius, 305-306

Palatine slavery, 14 Phaniscus (character in The Ghost), 12

Palestine, 175. See also Hebrews Pharaonic Egypt: internal source of slav-

Pallas, 301, 307 ery, 42: rituals of enslavement, 55;

Panama: treatment of slaves, 113; slave marks of servitude, 60; slaves as dowry,

population, 471n82 : 167; concubinal manumission, 230;

Papoulia, Basilike, 311-312, 449n34 freedman status, 253; prisoners of war, Paramoné (down payment), 238, 239, 282 403. See also Egypt Parasitism, see Human parasitism Philby, Harry St. John Briger, 92-93 Pardos (of mixed parentage), 269 Phillips, U. B., 335, 369n22 Parenthood, and manumission, 263-264 Pipes, Daniel, 314, 448n31, 449n32

Pascal, Blaise, 78 Piracy, see Kidnapping

Pastoral societies: master-slave relation- Pitt-Rivers, Julian, 79, 80 ship, 174-175, 180; manumission rate, Plantation slavery, 46; initiating rituals,

277 54; economic success, 93, 456n4; stereo-

Paternalism, 93-94, 339 types of slaves, 96; enslavement

Patrilineal societies: status inheritance methods, 116, 117; treatment of slaves, patterns, 137, 138, 139: peculium, 186: 173-174, 181, 190; manumission rate, freedmen, 241, 251; manumission rate, 284-285

279 Plato, 8; on recognition of power, 35; on

Paul (apostle), 70, 71, 306; on redeemed honor and pride, 81, 90; on inheritance man and God, 70, 76: on Christ’s cruci- status, 140; on timocratic character, fixion, 71, 74, 75; on manumission, 227 386n14

p,;;

Pawning of persons, 124, 125 Plautus, 12, 40, 381n93, 398n70 Peculium: defined, 182-184; disposal of, Pliny the Elder, 339 184-185, 210; recognition of, 185-186, Polanyi, Karl, 168

192: and crimes of slaves, 197: and level Polis (slaves), 66, 238

of skills, 254, 266, 267; prostitution and, Political eunuchism: power relationships,

263: and manumission rates, 270, 283 315; in Byzantium, 315, 317-319,

Peloponnesian wars. 114. 115. 288 321-330, 331; in China, 315-316, 318, Penal slavery: as source of slaves, 44, 327, in Persia, 31>, 316, in Turkey, 316, 126-129, 405n 146, 419n9; manumission ut Alrica, 3 16, sociological aspect.

and. 127: demand for labor and 319-320; in Tudor court, 330-331;

378n43 among the Yoruba, 451n72 Political manumission, 234-236, 287

Peons, 25, 26325, yo 327, way331, cult436n83 saves. a 322-323, Penal systems, 43 Pollution theory, with cult slaves Perak (Malaysia): slave raids, 118, Popo (West Africa), 123

400n85. debt-slavery, 125 Population decline, and slave trade, 157.

Perdue, Theda, 46-47 See also Slave populations Persian Gulf, 150 . Portuguese: kidnapping by. 117. 119:

Persians: enslavement patterns, 110; elite slave trade. 150, 160. 163: manumission

slaves, 299, 315, 316 rate, 289; “abolition” of slavery, 466n54 Persian wars, 115, 288 Postmortem manumission, 219-228, Personal characteristics, 85, 87; and mas- 433n45

ter-slave relationship, 176, 177. See also poteidaians (Greece), 111

Ethnic factors; Race Potestas (authority), 188, 189

Personal relations, levels, 12, 13 Potlatch ceremony, 191; honor aspect, 84, Peru: slave populations, 133; freedman 220; prestation in, 214; manumission

das, 427n73 380n79

status, 245, 255, 256, 289; Jesuit hacien- aspect, 222: assigning new name,

504 Index Poverty: and debt-servitude, 124-126; and Qoran, see Koran sale of children, 129, 130; and self- Quashee (slave term), 91, 338

enslavement, 130-131 Queensberry, Duchess of, 422n21

Power: relationships, 1, 2, 26, 307, 308,

332; honor and, 10-11, 13, 78, 79, 84, Rabinowitz, Jacob, 253 87; and property concept, 17-21; ritual- | Raboteau, Albert, 73, 74 ization, 19, 368n8; personalistic idiom, Race: and slave status, 58, 61, 254,

18-19, 27-28, 32, 33: materialistic 420n14; and range of color, 61; and idiom, 19, 28-32; origins, 26; need for master-slave relationship, 176-179; and recognition, 35; arrogance of, 36; sym- manumission rates, 264, 268, 272, 273, bols, 37; and total domination, 332-333 294; and recruitment of elite slaves,

Powerlessness, 2, 4, 5, 306 310; and status, 382n106, 421n19, Prestation, 212-213, 214, 215 441n18. See also Miscegenation

Prestige ranking, 247 Racism, 41, 177-178, 420n14 Price of slaves, 163-164, 166, 168, 283, Radbruch, Gustav, 128

441n13 Ransom, 107, 125, 130, 178, 380n79

Prisoners, see Criminals Rape, 82, 193, 206. See also Sexual exPrisoners of war: criminal status, 42, 43; ploitation and slave status, 43; as source of slaves, | Rapoport, Anatol, 336, 453n3 44, 106, 107-115, 119-120, 170, 400n90, Rattray, Robert S., 27-28, 136, 185 403n113; alternatives to enslavement, Rawick, George P., 3 106-107; treatment, 106, 108-109; ran- Raziya, Queen (of India), 422n20 som and exchange, 107; as source of in- Rebel, The (Camus), 204 come, 107-108; impressment, 109; in Rechtsverhaltnis (rights relationship), 36 kin-based societies, 113; in large-scale Reconciliation, 70 systems, 113-114; sexual bias, 121-122, Redemption, 70, 75; rituals, 214-219

in agriculture, 395n26. See also War- Red Sea area, 116, 150

fare Reenslavement: to a god, 72; and retalia-

Private slaves, 43 tion, 94; from freedman status, 241, Property: concept of, 17-21, 369n22, 245; economic conditions and, 287;

371n43, 373n54; Roman conception, 21, warfare and, 289, 292 28-32; Anglo-American conception, 21, Religion and symbolism: and intrusive

22, 369n22; and slavery, 21-27, 28, conception of slavery, 43-44, 45; in 369n21; legal aspects, 28-32; absolute, Slavery and manumission, 66-68; cults,

29, 32; feudal notions, 30 68-69; emperor worship, 69, 70; salva-

Prostitutes, 129, 400n85 tion and redemption, 70, 72, 75;

Prostitution, 177, 263, 370n29 Christ’s crucifixion, 71, 74, 75, 227.

Protestantism, 72, 73, 74 Catholic dualism, 72, 74, 75-76: converPubic hair, shorn, 60, 62 sion techniques, 72-73, 276-277; fundaPublic slaves, branding, 59 mentalism, 73~—75, 76, 260; in enslave-

Puerto Rico, 187 ment practices, 93, 107; and slave

Pulaaku (stereotype ideal), 84 marriage, 189; and killing of slaves, Pulleyblank, E. G., 86, 108, 127 192; and testamentary manumission,

Punic wars, 288 222, 224, 226; and freedman status, 260; Punishment: threat of, 3; forms, 59; and and manumission rate, 273-277, 282: in preservation of honor, 89-90, 128; and eunuchism, 323, 325; Exodus myth,

degradation, 93-94; enslavement as, 385n160

126-129; as social control, 206, 285; in- | Renegades, 277, 443n49

centives versus, 339-340; sale of slave Renner, Karl, 29

as, 419n4. See also Capital punish- Repp, R. C., 313

ment Reproduction: rates, 3; as wealth, 33; bio-

Purity and Danger (Douglas), 322 logical and social, 132-135, 170, 408n4

| Index 505

Residence of slave, and master-slave rela- slavery, 125; penal slavery, 126, 127,

tionship, 174-175 128; political manumission, 127, 219,

Reunion (Mascarene Islands), 257 235-236, 249, 253; child exposure, 130: Reward: reinforcement, 3; freedom as, slave populations, 133-134, 402n105, 260, 270, 340, 444n6S5; treatment as, 339 408n4; status inheritance pattern,

Rhodes (Greece), 152 , 139-141, 142, 143, 144, 146, 409n25, Rice, symbolism of, 216 411n68; slave trade, 152; internal trade, Richard II (character from Shakespeare), 164-165; slaves as dowry, 167; Greek

78 tutors, 174; slaves living apart, 175; ra-

Riesman, Paul, 83-84 cial differences, 177-178; skilled slaves, Rights, in jurisprudence, 20-21, 22 180; slave peculium, 184, 263; slave

Rigspula, \77 unions, 187, 189; crimes against slaves, Rigveda, 86 195: crimes of slaves, 196, 197, 241; Rio de Janeiro, 165 granting of sanctuary, 201; change of

Rites of passage, 53, 216 master, 202; freedman status, 216, 241, Ritual(s): of enslavement, 8-9, 38, 51-62; 242, 243, 248, 249, 252, 253, 305, 306,

as symbolic control, 37, 375n6; reli- 307-308; testamentary manumission, gious, 66-76; of initiation, 53-54, 191; 222-223, 224, 227, 238-239, 435n73: of gift exchange, 212-213, 236-237; concubinal manumission, 231; manumanumission, 214-219, 237; of adop- mission by adoption, 234; collusive

tion, 234 -manumission, 236, 435n71; reenslave-

Rodrigues (Mascarene Islands), 257 ment, 241; wala relationship, 242, 243, Roman Empire, 70, 76, 122, 144, 152, 224, 252, 253; incidence of manumission,

225 267, 269; military slaves, 288; familia

Roman law: property, 28-30, 32, 373n54; Caesaris, 299, 300-308, 319; upper role of slavery, 29; doctrine of domin- class, 304; imperial slaves, 332; latifunium, 30-32, 311; intrusive conception of dia slaves, see Latifundia slaves; paraslave, 40; asylum for slaves, 69; princi- sitic nature of slaveholding, 339; masple of privilege, 89, 195; exposure of terless slave, 366n13; sale of children, children, 130; status inheritance, 139; 370n27; racial prejudice, 420n14 marriage, 189; Twelve Tables, 195, 222; | Roman Spain, 288

postliminium, 215-216 Rope, The (Plautus), 12

Romans: social alienation of slaves, 7: Ross, Alf, 20, 21 conception of ownership and property. Rousseau, Jean Jacques, 2 21, 22, 28-30, 31-32, 371n4]1; sale of Rout, Leslie, 290 children, 24; nonslave persons, 24, 26, Royal families, 141

36; socioeconomic order, 28-30, Royal slaves, 195 372n49; and manumission, 30-31, 58, Ruffin, Thomas, 3-4 127, 210, 211, 215, 219, 229, 231: free Runaways, 59, 94, 270, 432n15, 447n97 laborers, 34; etymology of “slave,”’ 40; Runciman, Steven, 319

rituals of enslavement, 54, 55, 58, Rural slaves, 242; manumission rate,

380n75; Greek slaves, 55, 89, 90, 91, 92, 269-270, 282, 284-285 174; manumission rate, 58, 271, 274, Russell, J. C., 156-157 282, 283, 284, 288, 296; and religion, Russia: context of slavery, 39, 42, 44; 66-67, 68-70, 74, 76; nature of honor, means of enslavement, 42-43, 45: ritu77, 88-89, 92, 304, 306; value of pun- als of enslavement, 55; marks of serviishment, 89-90; attitude toward Greeks, tude, 59, 60; penal slavery, 128; self90, 91; enslavement of war captives, enslavement, 130; slave trade, 152; slave 107, 113, 396n52; enslavement of popu- peculium, 183; elite slaves, 299; human

lations, 111-112; enslavement by kid- sacrifice, 428n84 napping, 116, 121; killing of slaves, 121, | Russian Orthodox Church, 43-44 192; economy, 122-123, 403n114; debt- Rustici (agricultural slaves), 270

506 Index

Ryan, T. C., 224 Schternberg, Lev, 233

Rycaut, Paul, 308 Self-blame, 12

Self-defense, 200-202

Sab group (Somali outcastes), 49,50. 137, Self-enslavement, 130-131

195 Sellin, J. Thorsten, 127, 378n43

Sacral manumission, 237-238 Semenov, I. [., 109, 404n131 Sacrifice, see Human sacrifice Sena (Mozambique): concubinal manuSafawid dynasty (Persia), 316 mission, 228; sexual exploitation, 230; Sahara: master-slave relationship, 36; en- freedman status, 247 slavement of war captives, 114; slave Seneca, 92, 339 trade routes, 157-160; manumission, Senegambia (West Africa): enslavement

282 by warfare, 119; status inheritance pat-

Sahel, 252, 277, 282 tern, 147; slave trade, 160; manumis-

Sahlins, Marshall, 19, 28, 212, 213 sion, 268; slave population, 461n26. See Saint Domingue (Hispaniola), 249-250, also Wolof

275, 481 Separation, fear of, 6

Sale of children, 129, 370nn26,27 Sereer (Gambia): change of master, 202:

Sale of nonslave persons, 23-26 postmortem manumission, 222 Sale of slaves, 23, 175, 419n4. See also Serfs, 25, 26, 128, 411n68

Price of slaves; Slave trade Servitude: term, 7; hereditary, 9-10; volSalvation: religions of, 70; as reenslave- untary, 27-28. See also Debt-servitude ment to a god, 72; and manumission, Servus vicarius (slave of a slave), 184

225 Sex ratios: in slave populations, 199; and

Samas (sun-god), 218 freedman’s acceptance, 250; and manuSambo ideology, 91, 96, 207, 338 mission, 255

Sanctuary, 201-202 Sexual behavior: of wives of slaves, 6; and Sangir (Indonesia), 118 caste, 49; adultery, 128, 188; in master-

Santiago (Cape Verde), 274, 285 slave relationship, 173; legal sanctions, Santiago del Cobre (Cuba), 290 187-190; extramarital, 228; of eunuchs, 540 Paulo: demand for slaves, 165; manu- 326; of Muslim women, 440n3

mission rate, 290 Sexual bias: in enslavement practices,

S40 Tomé (West Africa): manumission 120-122, 129, 403n 108, 441n18; in rate, 274, 465n52; slave population, treatment of slaves, 179; in manumis-

465nn51,52,53 sion, 228-232, 239: and freedman staSaracens (Spain), 145, [52 tus, 240, 251 Sardinia, 152, 282 Sexual exploitation: of concubines, 50,

Sarhed (honor payment), 82 229, 230; in kin-based societies, 64-65,

Saria (holy law), 312 232; rape, 82, 193, 206; in developed

Saudi Arabia: honor in, 92-93; slave Slave systems, 85, 88; and marriage, trade, 159; change of master, 202-203, 189; legal restraints, 200; and manumis-

204. See also Arabs sion, 229, 230, 232: sanction of, 260-261

Sawyer, P. H., 154, 156 Sexual values, and freedman status, 259, Scandinavians: law and authority rela- 260, 261

tionships, 36; slave trade, 152-153, Seychelles (Indian Ocean), 257 154-156, 160; slave population, 157; Shakespeare, William, 78 slave stereotypes, 176-177; manumis- Sharp, William F., 273 sion ritual, 217; manumission by adop- Shaw, Stanford, 309 tion, 233; freedman status, 243-244, Shell, Robert, 268, 273, 276 245; manumission rate, 278; economic Sherbro (Sierre Leone): punishment of

role of slaves, 386n18 slaves, 128, 429n102; status inheritance

Schlaifer, Robert, 87 pattern, 146-147; tenant slaves, 181, Schrenck, Leopold von, 229-230 _ granting of sanctuary, 202

Index 507 Shild (Danish hero), 47 , Slave trade, 113; internal means, 118-119,

Shin, Susan, 42, 289 164-166; external means, 148-164;

Shorn head, 60-61, 128, 215 intra-African, 149, 156, 160, 399n78,

Siberia: penal slavery, 128; self-enslave- 403n113; medieval European, 149,

ment, 131. See also Gilyaks 152-157; volume. 150, 152, 156-157,

Sicilians (Italy): enslavement of, 111; en- 159, 160; Indian Ocean, 150-151; Black slavement by, 114; slave trade, 152; ten- Sea, 150-152, 153; Mediterranean, ant slaves, 181; manumission rate, 274, 150-152, 154, 157-158, 160, 170-171:

276 routes, 152-154, 157-158, 159: Saharan,

Sidam province (Ethiopia), 316 157-159, 160; Atlantic, 159-164, 165,

Siegel, Bernard J., 188, 253 286; mortality rates, 159, 162, 163; Sierre Leone, see Sherbro demography, 160, 162; price of slaves,

Sirqu (Babylonia), 145 163-164, 441n11; interregional, 165, Sjaelland (Denmark), 244 166, 286; abolition, 165, 286

Skills, level of: and master-slave relation- Slavs: Viking raids on, 116; self-enslave-

ship, 179-180; and freedman status, ment, 131; trading, 155, 156; origin 254-255: and manumission, 264—266, of term, 250, 406n172; manumission,

267, 269, 270, 439n39 268

Skin color, see Race Smith, M. G., 41

Slave: origin of term, 40, 156, 250, Snowden, Frank M., Jr., 177 406n172; stereotypes, 338; origin of Social alienation, 7

concept, 376n17 Social death: conceptions, 38; intrusive

Slave breeding, 427n73 and extrusive modes, 39-45; as liminal Slave markets, 40, 459n12 state, 46; irrevocability, 215 Slave populations: mass enslavement, Social relations, of slaves, 6 110-112; biological reproduction, 132, Sodomy, 263 408n4, 414n30; social reproduction, 132; Sokolowski, F., 237

birth/death rates, 132, 133; decline, Sokoto (Nigeria): enslavement for tribute, 133; and demand for slaves, 157; size, 123; treatment of slaves, 199: manumisand treatment of slaves, 180, 198-199, sion rate, 274 206; sex ratios, 199; type, and manu- Solon, 125 mission, 255; and freedman status, Somali (Somalia): occupational castes, 49; 259-260; of large-scale slave systems, marks of servitude, 60; honor among,

353-364 84; status inheritance pattern, 137-138,

Slave raids, 281, 282, 399nn75,78, 400n85 141, 409n25, 410n27; absentee slave

Slave revolts, 259-260, 291, 465n52 owners, 180; slave unions, 188; crimes Slavery: constituent elements, 1-14; defin- against slaves, 195: manumission pating, 13, 21, 369n18, 371n39; internal re- terns, 228, 235, 249; freedman status, lations, 17-101, 296; ownership concept, 249; manumission rate, 271

21-27; idioms of power and, 27-32; Somaliland (East Africa), 150 contradictions, 32—34, 296: terms for, Somatic theory of manumission, 268-269 40; and culture, 84-85; and timocracy, Sorel, Georges, 3 94-95; dialectics, 97-101; institutional Sosia (character in Amphitryo), 12 process, 104-171; versus debt-servitude, South Africa: enslavement methods, 111, 124: condition of, 172-208; as relation 117, 119; status inheritance patterns, of domination, 334-335: as relation of 138; slave trade, 160; tenant slaves, 181.

parasitism, 335-336 manumission, 231, 232; freedman staSlave status: need for new persons, 3; out- tus, 245, 246, 254, 255, 256, 439nS0; sider, 7; inheritability, 9; with adoption, architecture, 255; racism, 256, 383n107; 233; permanent, 280, 404n 134; in war- incidence of manumission, 263, 266, fare, 291-292; ambiguity, 425n44. See 268, 269; manumission rate, 273, 276,

also Status inheritance patterns 285

508 Index South America: marks of servitude, 60; Stigma, of slavery, 247, 251, 252, 253, 257, absenteeism, 423n33. See also names of 294

individual countries Structuralism, 323

South Carolina: treatment of slaves, 6; Struve, V. V., 109 master-slave relationship, 12; slave Submission, 77, 78, 95 naming rituals, 56, 57, 58, 380n80; Sub-Saharan Africa: nature of honor, 82; branding of slaves, 59; status inheri- enslavement of war captives, 114; incitance pattern, 144; freedman status, dence of manumission, 268. See also

246; regulation of clothing, 381n91; South Africa

slave population, 475n141, 483. See also Subsistence mode: and treatment of

U.S. South Slaves, 181; and peculium, 186

Southeast Asia: caste in, 49: enslavement Sudan: marks of servitude, 58: internal

by kidnapping, 117, 399n75 kidnapping, 117, 400n90; slave trade, Spain: enslavement of Europeans, 44; en- 149, 169; treatment of slaves, 175; slavement of war captives, 114; piracy, freedman status, 252; manumission

116; penal slavery, 128; status inheri- rate, 277, 282 tance patterns, 141, 144, 145-146; slave | Suetonius, 307 trade, 152, 154, 160; tenant slaves, 181; Sugar plantations, 116, 117

slave peculium, 184, 192-193; killing of Sulayman Kean (sultan), 313 slaves, 192; crimes against slaves, 196; umer, - See also Mesopotamia crimes of slaves, 197; change of master, Surinam (South America): freedman sta-

202: contractual manumission, 239. tus, 255, 257, manumission rate, 275, freedman status, 243; manumission 278, 285 » 287

, wansea (Wales),

rate, 270, 274, 283, 284, 288 suzukh P Wa les 54 Spanish Caribbean: decline of slave popu- Sweden: status inheritance pattern, 140

lation, 113; enslavement methods, 114, ; P , , 120: manumission. 290 145; slave trade, 154, 156; testamentary Sparta: e , ; . enslavement manumission, 226; concubinal manuAthenians, | 11; —. mission, 230; of manumission rate, 276; helotry, 112; timocracy, 386n14. See . also Greeks317, S ——326 Spores,

S

economic role of slaves, 386n18

ymbiosis, 335, 453n1

; ymbolic control, authority as, 36-38

Sports, professional, 24-26, 371n35 Syria, 316

Spranger, Peter P., 398n70 Syrus, Pubilius, 77-78 Starkey, David, 330, 331

Starr, Chester G., 89 Taboos, 129 :

Statistical methods, 345-349 Tacitus, 305, 306

Status, see Freedman Status; Slave status Tallensi (Ghana): fictive kinship system, Status inheritance patterns, 113, 170, and 63: manumission rate, 271 social reproduction patterns, 134—135, Tandojai (guardian of the dead), 221 170; Ashanti pattern, 135-137; of non- T’ang dynasiy (China), 315 Slaves, 136; Somali pattern, 137-138, Tanzania, see Kerebe 141, 409n25, 410n27; Tuareg pattern, Tapu (Maori laws), 51

136, 138-139, 145; Roman pattern, Tattoos, 58-59

139-141, 143, 146, 409n25, 41 1n68; Tax payment, enslavement for, 122Chinese pattern, 141-144, 145, 409n25: 125 Near East pattern, 144-146; Sherbro Temple slavery, 49, 237-238 pattern, 146-147; with adoption, 233; Tenant farming, 181, 199, 425n41 race and, 277-278; ghilman and, 312. Terence, 40

See also Hereditary slavery Testamentary manumission, 237, 239,

Stearns, Robert E., 84 435n73; origins, 222-223; salvation and,

Steiner, Franz, 19 225-226, 227

Index 509 Thailand: enslavement by kidnapping, ethnics, 422n23; slave population, 117, 399n75; debt-slavery, 125; king’s 461n27 slaves, 174; killing of slaves, 191-192 Tuden, Arthur, 147 Theodore (archbishop of Canterbury), Tudor Court, 330-331

192 Tung Shu-yeh, 108 _

Third Dynasty of Ur: slaves as foreigners, Tunis, 277, 443n49 40; sources of slaves, 44; ransom and Tunisia, 157 warfare, 107; slave unions, 188; freed- Tupinamba (Brazil): rituals of enslaveman Status, 253. See also Mesopotamia ment, 52; nature of honor, 81; treat-

Thirteenth Amendment, 25 ment of prisoners, 106-107, 402n102;

Thralldom, 81, 82, 177 manumission rate, 27] Thucydides, 35 Turks: enslavement for tax payment, 123; Timocracy, 94-97, 100; derivation, 386n14 self-enslavement, 131; slave trade, 152; Tiv (Nigeria): rituals of enslavement, 53; racial differences, 178; incidence of

manumission rate, 271 manumission, 268, 283; eunuchs and

Tlingit (northwest coast of America): Slaves, 316, 332. See also Mamluks marks of servitude, 58; potlatch cere- Turner, Victor, 37 mony, 84, 191, 222; status inheritance Tuscany (Italy): social death, 39; status pattern, 136; skills, and treatment, 179; inheritance pattern, 144; granting of sacrifice of slaves, 191, 222; manumis- sanctuary, 202 sion, 222; source of slaves, 412n6 Tutors, Greek, 91, 174 To Anda’e tribe (Toradja group), 85 Twelve Tables, Laws of, 195, 222

To Lage tribe (Toradja group), 85, 144 Twins, 129 To Pebato (Toradja group), 85 Toradja (Celebes): rituals of enslavement, | Ugaritic manumission, 218

54; nature of honor, 84-85; sources of Ui Neill ({rish royal family), 141 slaves, 113; status inheritance pattern, Ukodei (adopted slave), 233 144; treatment of slaves, 176; change of | Uktena (mythical beast), 47 master, 203; manumission, 221, 271: Unruh, Ellen, 183 sacrifice of slaves, 221; manumission Ur, see Third Dynasty of Ur rate, 274; parasitism, 337; origins, Urban slaves, 254; manumission rate,

396n46 269-270, 282; economics, 456n5

Towara (Bedouin tribe), 194 U.S. South: incidence of manumission, 3, Trachalio (character in The Rope), 12 267; sexual unions and marriage, 6,

Tradition, and law, 36-37 — 190; black and white servitude, 6-7: Transoxiania (West Asia), 310 economics of slavery, 11, 33, 65, 190; Treatment of slaves, see Conditions of servile personality, 12, 367n41; free la-

slavery borers, 34, 374n61; nonslaveholders, 36;

Treggiari, Susan, 174, 242, 243, 252, 304 caste concept, 48; surnames of slaves,

Tribute, enslavement as, [22-124 56; kin terms, 65; religion and slavery,

Trimalchio, 305-306 73-75, 260; free persons and slaves, 88,

Tripoli (North Africa), 277 94; honor and slavery, 94-97, 367n41; Trumai Indians (Brazil), 402n102 European influence, 95; Sambo ideolTsimshian (northwest coast of America), ogy, 96, 207, 338; slave trade, 160; in-

136 ternal slave trade, 165-166; treatment

Ts’ui Tao-ku, 252 of slaves, 206-208; manumission rate,

Tuaregs (Algeria): master-slave relation- 217, 255, 273, 284, 285, 286, 292-293, ship, 4, 46; status inheritance pattern, 295: manumission as control, 217, 232, 136, 138-139, 145; bridewealth, 167; 285, 444n65; wala relationship, 245, racism, 176, 178-179; tenant slaves, [81; 246, 260; selective manumission, change of master, 203; manumission, 246-247, freedman status, 255, 256, 271, 437n23: enslavement of fellow 259, 260, 261; hostility to manumission,

S10 Index

U.S. South (continued) Volga River, 154 255, 259; artisans, 255, 256; miscegena- Voluntary servitude, 27-28 tion, 255, 261; demographic mix, 259;

exploitation of slave women, 260, 261; Wadai (Chad), 124 abolition of slavery, 285, 286; military Wagner, Edward W., 42 manumission, 292-293: “achievements” Wala relationship, 241-247, 251, 294; and of slavery, 335; parasitism, 336-337; freedman status, 243, 244, 250, 253, slave populations, 414n30, 475nn 139. 257, 311; intersocietal variations, 244; 141, 483. See also names of individual legal enforcement, 244, 252; absence,

states 260

Wali (Aswan), 124

Vai (West Africa): sexual bias in enslave- Wallon, Henri, 4-5 ment, 122: internal warfare. 178; manu- Wang Yi-T’ung, 143

mission pattern, 280 Warfare: as source of slaves, [06-1 IS, . Valencia (Spain), 263 119-120, 170, 396n52, 397n54; logistics,

Valgard, 154 , 106; for wife recruiting, 107, 402n 102; Van Gennep, Arnold, 37, 322 atrocities, 120, 212; manumission and, Varro, 86, [87 287, 289-293; slaves in, 290-291. See Vaughn, James H., 45-46, 137, 378n55_ also Prisoners of war Vebjorn (character in Icelandic saga), 79 Water, symbolism, 216, 218

Vediovis, cult of, 223 Watson, Alan, 366n13, 373n54

Venezuela, see Goajiros Watson, James L., 21, 296, 370n26

Venice, 296 Wealth, 2, 28, 33, 371n4l

Verbal abuse. 193 Weaver, P. R. C., 40, 300, 303-304

Verlind Ch les. 192. 270. erlinden, Charles, , ; 276 Weber, Max, 1, 2, 36-37, 375n6, 384n138 Vespasian (emperor of Rome), 307 . Welsh, nature of honor, 82

V; oopsslavery Welskopf, Elisabeth,Ww 3, 303 ietnam: versus debt-servitude, 86, 125, 126; killing of slaves, 191-192: ergeland, Agnes, 217, 218 , 8 oo - West200; Africa: slaves asmanumiswealth, 33;OA slave self-defense, concubinal .sion,trade, 114, 119, 128, 148-149, 156, 157, 230; freedman status, 252 160, 164, 400n90; selection of slaves

Vikings enslavement , , i. Ing UrOP(Europe): y 122; penal slavery, 128; slaveby peculium, kidnapping, 116; slave trade, I52- 185: freedman status, 443n53

153, 154-155; ships, 156; killing of Westermann, William, 189, 398n70 slaves, 191: economic role of slaves, Westermarck, Edward, 369n18

386n18 a West Indies, see Caribbean

Violence: and domination, 2, 3; and Westington, Mars M., 121 master-slave relationship, 3-4; in self- Westrup, C. W., 31

hatred, 12 | Whipping, 3-4, 12, 56, 94, 128

Virginia: indentured servants versus Wife recruiting, 107, 402n102 slaves, 9; status inheritance pattern, Wilbur, C. M.. 108, 127 138; manumission, 292: runaways, Wilks, Ivor, 479n90 447n97; slave population, 475nn1t39, Williams, Carl O., 81

141, 483. See also U.S. South Wilson, David, 200

Virtus (moral integrity), 304-305 Winnebago Indians (North America), 325 Visigothic Spain: status inheritance pat- Witchcraft, 64, 327 tern, 141, 144, 145-146; slave trade, Wittek, Paul, 312

152; slave estates, 181; privileged slaves, Wittfogel, Karl L., 315, 319

196; crimes of slaves, 197; freedman Wodaabe Fulani (Niger), 271 status, 243; incidence of manumission, Wolof (Gambia): status inheritance pat-

270; manumission rate, 274, 288 tern, 147; bridewealth, 167; change of Vitellius (emperor of Rome), 307 master, 202: sacrifice of slaves, 222, 228:

Vogt, Joseph, 87 manumission rate, 271

Index 511 Women: in U.S. South, 261; in Byzan- Yoruba (Nigeria): marks of servitude, 59; tium, 325, 326, 329; in China, 325, enslavement, 120; slaves as money, 168;

326 in Cuba, political eunuchism, Wood, Peter,401n97; 56 451n72 Woodhead, A. Geoffrey, 35 Yuan period (China), 142 Worthy Park plantation (Jamaica), 6, 57, 59 Zaire, see Kongo; Nkundu Wrigley, Philip K., 25-26 Zambia, see Bemba; Lozi; Luvale Zandj (Iraq): freedman status, 248, 250,

Xurasanians (free Iranians), 310 299, marginality, 332; stereotype, 299, 338; revolt, 467n63

Yangban (aristocrats), 86 Zanzibar: inferior slave status, 93, 94: enYao (East Africa): stigma of slavery, 251; slavement by kidnapping, 117; slave

slave status, 280, 464n45 trade, 150; manumission rate, 274; elite

Yi dynasty (Korea): context of slavery, slaves, 316 39, 42; self-enslavement, 130; status in- Zaria (Nigeria), 461n26

heritance pattern, 143; slave estates, Zimmern, A., 283 181; slave population, 469n73